Skip to main content

Full text of "Color removal from a neutral sulfite waste using magnesium coagulation"

See other formats


COLOR REMOVAL FROM A NEUTRAL SULFITE 
WASTE USING MAGNESIUIvl COAGULATION 



by 

JAMES S. TAYLOR 



A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF 

THE UNIVERSITY OF FLORIDA 
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE 
DEGREE OF DOCTOR OF PHILOSOPHY 

UNIVERSITY OF FLORIDA 
1976 



TO 

JANET 

JIMMY 

o 

AND 

BRIT 



Acknowledgements 

I wish to express my gratitude to my committee chairman, John 
Zoltek, Jr. for his overall guidance, understanding and friendship 
in assisting me in my research. I am deeply indebted to T. deS 
Furman and J. Edward Singley whose technical and exemplary contri- 
butions to my education will never be forgotten. The technical 
insight and timely assistance given me by Ellis D. Verink, Jr. are' 
sincerely appreciated. I wish to extend my appreciation to H.F. 
Berger, who, through the National Council for Air and Stream Improve- 
ment, made the funding of my research possible. I will always hold 
his cooperation and patience in high esteem. The contributed research 
and extensive laboratory work by Gary Christopher and Bevin Beaudet in 
completing their masters projects is acknowledged and appreciated. 

I am sincerely grateful for the sacrifices my wife Janet made 
and the contributions of my parents in enabling me to pursue my 
education. The values set forth by my parents years ago came to 
bear during my research. 

I have had many rewarding experiences at the University of 
Florida and am grateful for the opportunity to have been part of 
that institution. 



1X1 



Table of Contents 



Page 



111 



Acknowledgements 

List o£ Tables ^^^ 

List of Figures 

Abstract 

Chapter 



IX 

xii 



1- INTRODUCTION • 1 
1-1 General Background 1 
1-2 Legal Requirements 3 
1-3 Purpose of This Work 3 

2- COLOR 5 
2-1 Color in the Electromagnetic Spectrum 5 
2-2 Lignin 7 
2-3 Characteristics of Color 9 
2-4 Coagulation 13 
2-5 Color Removal by Coagulation 23 
2-6 Magnesium Coagulation 25 
2-7 Color in Pulp Mill Effluents 27 

3- LABORATORY PROCEDURES 32 
3-1 Feed Solutions 32 

3-1.1 Synthetic Waste Solutions 32 

3-1.2 Coagulation Chemicals 33 

3-1.3 Polymers 33 

3-2 Analytical Equipment and Techniques 34 

3-2.1 Total Carbon Measurements 34 

3-2.2 Color Measurement 34 

3-2.3 Incineration 35 

3-2.4 Jar Tests 35 

3-2.5 Metal Analysis 37 

3-2.6 Mobility Measurements 37 

3-2.7 pH Measurements 38 

3-2.8 Settling Tests 38 

3-2.9 Solids Analysis 38 

3-2.10 Titration Curves 39 

3-3 Experimentation 39 

3-3.1 Coagulation Experiments 39 

3-3.2 Coagulant Recovery ■ . 39 
3-3.3 Coagulant Recycle 



40 



lY 



Chapter ^^g® 

4- RESULTS ^^ 
4-1 Determination of Coagulation pH and 

Coagulant Dose ^^ 

4-1.1 Coagulation pH ^1 

4-1.2 Coagulant Dose 43 
4-1.3 Variation of Coagulation pH with 

Coagulant Dose 4.4 
4-1.4 Magnesium Remaining in Solution as 

a Function of Final pH 47 
4-1.5 Magnesium and Ca[OH)„ Dose as a 

Function of Initial Waste Color 49 

4-2 Waste Characceristics 51 

4-2.1 Untreated Waste Titration Curves 5,1 
4-2.2 Comparison of Untreated and Treated 

Waste Titration Curves 58 

4-2.3 Waste Content ■ 60 

4-3 Color Removal Mechanism 60 

4-3.1 Color and Magnesium Titration Curves 60 

4-3.2 Magnesium, Calcium, Color and Organic 

Carbon Residuals After Coagulation 63 
4-3.3 Stoichiometry of Color Removal from 

NSSC Waste hy Magnesium Coagulation 71 

4-4 Settling of Coagulated Wastes 75 

4-4.1 Purpose of Settling Tests 75 

4-4.2 Sludge Settleability 77 

4-4.3 Mechanisms of Sedimentation 82 

4-5 Magnesium Recovery and Recycle 89 

4-5.1 Recovery Methods 89 

4-5.2 Process Reversibility 90 

4-5.3 Color-Cation Interaction ^^ 90 

4-5.4 Chemical Equilibrium of Mg -CO2-H2O 96 

4-5.5 Sludge Incineration 103 

4-5.6 Magnesium Recovery 107 

4-5.7 Magnesium Reuse 118 

5- DESIGN OF A COLOR REMOVAL PROCESS FOR A NSSC 

WASTE USING MAGNESIUM COAGULATION AND RECOVERY 123 

5-1 Coagulation 123 

5-2 Sedimentation • 124 

5-3 Vacuum Filtration 127 

5-4 Incineration 129 

5-5 Carbonation 130 

6- COST- 134 
6-1 Chemical Costs 134 
6-2 Capital and Operation Costs ' 134 
6-3 System Costs 137 



Chapter P^S® 

7- CONCLUSIONS AND RECOMMENDATIONS 145 

7-1 Conclusions 145 

7-2 Reconmendations 147 

REFERENCES 149 

Biographical Sketch 156 



vi 



List of Tables 

Table Title Page 

2-1 VISIBLE SPECTRUM AND COMPLIMENTARY COLORS 6 

4-1 GRAPHIC DETERMINATION OF pK^ OF SODIUM BASE NSSC WASTE 57 

4-2 UNTREATED AND TREATED NSSC WASTE ANALYSIS 61 

4-3 POLYMER DESCRIPTION AND SVI FOR POLYMER ASSISTED 

SLUDGES 78 

4-4 ELECTROMOBILITY AND ZETA POTENTIAL FOR MAGNESIUM 
SLUDGE PRODUCED IN TAP WATER AND NSSC WASTE AT 
•VARYING pH 86 

4-5 CaCO. PRECIPITATION IN A NSSC WASTE 94 

4-6 CHEMICAL REACTION AND pK VALUES CONSIDERED FOR 

Mg"*'"^-C02-H20 SYSTEM 100 

4-7 AVERAGE CHARACTERISTICS OF A SLUDGE PREPARED 

BY COAGULATING A NSSC WASTE 104 

4-8 MgO REACTIVITY AS AFFECTED BY TEMPERATURE 106 

4-9 INCINERATED SOLIDS ANALYSIS 108 

4-10 CARBONATION OF INCINERATED SLUDGE AT VARYING 
CONCENTRATIONS OF NONVOLATILE SOLIDS FOR 
MAGNESIUM RECOVERY 113 

4-11 COLOR REMOVAL BY LIME-MAGNESIUM COAGULATION 

USING THE SAME MAGNESIUM THREE TIMES 119 

4-12 COLOR REMOVAL BY LIME -MAGNESIUM COAGULATION 

USING THE SAME MAGNESIUM TWICE 121 

5-1 SOLIDS LOADING FROM SETTLING BASIN 128 

5-2 DESIGN SUMMARY FOR THE TREATMENT OF A NSSC WASTE 132 

6-1 • CHEMICAL COST TO TREAT A NSSC WASTE 135 



Vll 



Table Title Page 



6-2 UNIT OPERATIONS COST SUMMARY 

NSSC WASTE COLOR = 5000 139 

6-3 PROCESS COST SUMMARY. IN $/1000 GALLONS OF 

NSSC WASTE 140 

6-4 UNIT OPERATION COST SUMMARY 

NSSC WASTE COLOR = 2500 142 

6-5 NSSC PRODUCT COST INCREASE DUE TO COLOR 

REMOVAL BY MAGNESIUIvl COAGULATION 144 



Vlll 



List of Figures 



Figure Title 

1.1 NSSC flow diagram 

2.1 Quinonemethide 

2.2 Constitution scheme for lignin 
3.1 Standard Pt-Co color curve 

9 

4.1 Color residual as a function of final pH 

4.2 Color residual as a function of final pH 

4.3 Comparing NaOH and Ca(0H)2 for color removal 
via magnesium coagulation 

4.4 Comparing NaOH and CaCOH)^ for color removal 
via magnesitim coagulation 

4.5 Verification of coagulation pH' 

4.6 Magnesium remaining in solution as a function 
of final pH 

4.7 Lime dose as a function of initial waste color 
for magnesium coagulation 

4.8 Magnesium dose as a function of initial waste 
color using lime 

4.9 Titration curve of sodiiom base Mead NSSC waste 
with color equal to 2500 

4.10 Titration curve of sodium base Mead NSSC waste 
with color equal to 5000 

4.11 Titration curve of sodium base Mead NSSC waste 
with color equal to 10,000 

4.12 Titration curve of sodium base Mead NSSC waste 
with color equal to 20,000 

4.13 Titration curve of sodium base Mead NSSC waste 
with color equal to 40,000 



Page 

2 

8 

10 

36 

42 
42 

45 

45 
46 

48 

50 

50 

52 

53 

54 

55 

56 



IX 



Figure 



Title Page 



4.14 Titration curve of treated and untreated 

NSSC waste ^^ 



4.24 SVI and zeta potential vs. polymer dose for 
an anionic polymer #837A 



62 



64 



64 



68 



4.15 Titration curve of raw waste dosed with 
magnesixim 

4.16 Organic carbon and color residuals as a 
function of final pH 

4.17 Magnesium and calcium residuals as a function 
of final pH 

4.18 Color, T.O.C., and Mg"^ residual after Mg 
coagulation using Ca(0H)2 for pH control 

4.19 Color, T.O.C., and Mg"^"^ residuals after Mg 

• coagulation using NaOH for pH control 69 

4.20 Ratios of [OHJ^, ++/ Mg*"^ 72 

'- -"Mg 

4.21 Color and pH of a NSSC waste as a function 

of CaCOH) concentration ^6 

4.22 Sludge settling velocity for polymer assisted 

and raw sludge ^■'• 

4.23 SVI and zeta potential vs. polymer dose for 

a nonionic polymer #1905N 85 



85 



4.25 Zeta potential of magnesium solids in tap 

water and NSSC waste at varying pH 87 

4.26 Equilibrium concentrations of Mg and MgCOH^ 

with Mg(0H)2 at varying pH 87 

4.27 Color reversibility bar graph 91 

4.28 Color remaining as a function of CaCO 

precipitation • ^-^ 

4.29 Color remaining as a function of MgF2 

precipitation "^ 

4.30 Activity ratio diagram for log C = -1 97 



Figure Title Page 

4.31 Solubility diagram of Mg"^"^ in a C-j. = lO""^ M 

carbonate system 98 

4.32 Predominance diagram for log Mg = -1 99 

4.33 Color/Mg'^''' ratio as a function of incineration 
temperature 109 

4.34 Precipitation of MgCO^ -3^120 by aeration at 

various temperatures • 111 

4.35 pH as a function of carbonation at various 

nonvolatile solids concentrations 114 

4.36 Magnesium recovered as a function of 

carbonation time 115 

4.37 % magnesium recovery as a function of 

nonvolatile solids concentration 116 

5.1 Design data for sedimentation 126 

5.2 Flow diagram for lime -magnesium color 

removal process 133 



XI 



Abstract of Dissertation Presented to the 
Graduate Council of the University of Florida 
in Partial Fulfillment of the Requirements for the 
Degree of Doctor of Philosophy 



COLOR REMOVAL FROM A NEUTRAL SULFITE WASTE 
USING MAGNESIUM COAGULATION 

■ By 

James S. Taylor 

August 1976 

Chairman: John Zoltek, Jr. 

Major Department: Environmental Engineering Sciences 

A color removal process was developed on a laboratory scale that 
would remove 90% of the initial color of a neutral sulfite semi-chemi- 
cal (NSSC) pulp waste. The colored waste was coagulated at a pH of 
11, with stoichiometric amounts of magnesium and Ca(OH) . The magne- 
sium and Ca(OH) doses were represented by linear equations. 

The amount of magnesium required for 90% color removal was reduced 
25% when Ca(OH) was used for pH control. The reduction in coagulant 
dose was due to the chelation of the divalent calcium ion and organic 
acids in the waste. Titrametric techniques demonstrated that the color 
removal process removed 40% of the acid strength of the NSSC waste, and 
that 65% of the acids removed had a pK greater than 9. 

The zeta potential of the coagulated NSSC waste was -1.00 mv at 

++ 
pH 10.3 and zero at pH 12.5. Measurements of the (Mg ), [OH ), orga- 
nic carbon concentration and color removed during the coagulation pro- 

++ 
cess indicated that the Mg ion first chelated the organic acids 



Xll 



causing a 35% color increase. The Mg ion then formed a precipitate 
which resulted in color removal . The empirical formula for the pre- 
cipitate was Mg(OH) R, where R represents the precipitated organic 
acids. Once the magnesium precipitate formed, the molar ratios of 
the magnesium removed to the hydroxides removed was 1.5 for varying 
magnesium doses at constant pH. The consistency of the molar ratio 
at varying doses indicated color bodies were removed in magnesium 
coagulation by a chemical reaction. 

The color removal process was demonstrated to be completely 
reversible by varying the pH. In order to reuse the magnesium, the 
sludge was incinerated to remove the color from the magnesium solids. 
The optimum temperature of incineration was found to be 550 C. After 
incineration, all of the magnesium was recovered by bubbling a 10% 
CO^-90% air mixture through a slurry containing an incinerated solids 
concentration of 5318 mg/1. The fraction of magnesium solubilized 
from the incinerated solids was controlled by the Mg -CO^-H„0 system. 
The controlling solid phase was MgC0_'3H^0. 

The same magnesium was used three times to remove 90% of the 
color from three separate aliquots of NSSC waste. After three uses 
of the coagulant, 93% of the magnesium was recovered. 

The cost of using this process to treat a NSSC waste with an 
initial color of 2500 and a flow of 10 mgd was estimated to be 
$0.27/1000 gal. 



Xlll 



CHAPTER 1 



INTRODUCTION 



1-1 General Background 

Pulp and paper manufacturing is one of the largest industries 
in the United States. It is also one of the^ajor water using indus- 
tries in the nation, producing from some mills extremely large vol- 
umes of highly colored effluents", which are typically discharged to 
waterways. Color creates a unique problem in a stream. It is readily 
identifiable in an aesthetic sense and can detract from the natural 
beauty of a body of water. The amount of light penetrating a stream 
would be affected by a colored waste discharged to that stream, and 
could threaten the eco-system in that stream. The National Council 
for Air and Stream Improvement, a pollution abatement research organi- 
zation sponsored by member pulp and paper companies, has recognized 
this problem and has sought for many years to devise economical and 
effective color removal processes for all pulp and paper plant 
effluents. 

A highly simplified NSSC pulping process diagram is illustrated 
in Figure 1.1. The wood is prepared for the digestion process by 
removing the bark and increasing the surface area by a chipping pro- 
cess. The wood chips are then screened, fed into a digestor and 
mixed with a sulfite cooking liquor. The function of the sulfite 



Sulfite 

Cooking 

Liquor 



Water 



- 




Debarking 
Chipping 

Screening 






■ 
1 


' 






Digestion 












' 


1 






Washing 












■ 


t 






Bleaching 
(chlorine) 
(oxygen) 






< 


' 






Washina 













I 



Colored 

Waste 

Effulent 



Further 
Processing 



Fig. I.I NSSC flow diagram 



cooking liquor is to separate the lignin from the wood fiber. After 
the digestion process, the highly colored water soluble lignin is 
separated from the pulp by washing. The aqueous washings constitute 
part of the waste effluent. Depending on the ultimate use of the 
pulp, additional color removal may occur. Bleaching will further 
lighten the pulp. After each bleaching operation the pulp is washed, 
producing additional color in the final waste effluent. 

1-2 Legal Requirements 
In 1968 the United States Government passed the Clean Water 
Act. It, was amended in 1972 to include all pulp and paper mills 
using a NSSC production process. In the Federal Register under Pulp, 
Paper, and Paperboard Point Source Category, Effluent Guidelines and 
Standards, this law states in summary: All NSSC plants must remove 
75% of their effluent color by 1983, and all new NSSC paints built 
after 1975 must remove 75% of their effluent color. 

1-3 Purpose of This Work 
It was the purpose of this research to develop a color removal 
process for NSSC waste and to give insight into the mechanism by 
which that color was removed. The first objective was to develop 
a method of NSSC color removal that could be evaluated for use as 
a full-scale treatment process. The investigation was limited to 
jar testing- techniques, with subsequent sludge incineration and 
coagulant recovery on a laboratory scale. The second objective was 
to investigate the mechanism by which color removal occurred. 



Techniques employed in this phase of the research were chemical 
analyses in conjunction with the determination of the stoichiometric 
relationships developed in the color removal chemical reactions. 



CHAPTER 2 



COLOR 



2-1 Color in the Electromagn.etic Spectrum 
Color is a qualitative parameter that does not lend itself to 
exact engineering measurement. Within the visible region of the 
spectrum, persons with normal color vision are able to correlate the 
wavelength of light striking the eye with the subjective sensation 
of color. Table 2-1 shows the color perceived related to the wave- 
length. 

Objects are seen by either transmitted or reflected light. When 
"white light," containing the entire spectrum of visible wavelengths, 
passes through a medium such as a solution of NSSC waste, the medium 
appears colored to the observer. Since only the transmitted waves 
reach the observer, their wavelengths determine the color of the 
medium. Chromophores , or color producing compounds, absorb certain 
wavelengths of the spectrum depending on the electronic structure of 
the compound. A change could occur in the electronic configuration 
of a compound which could change it from a coloirless to a colored 
compound. The oxidation of an alpha-quinone would produce a colorless 
degradation product, but the self -condensation of the same alpha-qui- 
nbne would produce a colored product. 

Very little evidence has been gathered on the amount of environ- 
mental degradation caused by color. Properties of pollutants such 



TABLE 2-1 
VISIBLE SPECTRUM AND COMPLIMENTARY COLORS 



Wavelength, mu 



400 - 


- 435 


455 - 


■ 480 


480 - 


- 490 


490 - 


- 500 


500 ■ 


-560 


560 - 


- 580 


580 ■ 


- 595 


595 ■ 


- 610 


610 ■ 


- 750 



Color 



Violet 

Blue 

Green-blue 

Blue-green 

Green 

Yellow-green 

Yellow 

Orange 

Red 



Complimentary Color 

Yellow-green 

Yellow 

Orange 

Red 

Purple 

Violet 

Blue 

Green-blue 

Blue-green 



Source: Day, R.A., Jr., Underwood, A.L., Quantitative Analysis , Second 
Edition, Prentice-Hall, Englewood Cliffs, N.J., (1967). 



as available nutrients or oxygen demand have been shown to degrade 
the environment. However, the discharge of highly colored effluent 
would definitely affect the aesthetic quality of the receiving waters. 
Color would have a detrimental effect on process water used in the 
production of highly bleached paper. 

2-2 Lignin 

Lignin is one of the most abundant natural products on earth, 
constituting about one-fourth of the woody tissue in plants. It 
is responsible for most of the color present in natural waters. The 
natural formation of this cross-linked polymeric material from coni- 
feryl alcohol and related substances is not presently completely 
understood. Despite considerable research, the structural characteri- 
zation of lignin has been only partially successful. 

Freudenberg (1966) gathered information about lignin structure 
from direct oxidation of lignin, from bio-chemical experiments related 
to alcohols, and from lignin degradation with strong alkali, methyla- 
tion and oxidation. His experiments enabled an estimation of the 
relative amount of alcohols which served as building blocks of lignin. 
Lignification occurs in plant cells when alcohols are liberated and 
oxidized by natural organic compounds in the presence of air. The 
free radicals produced then combine and build up lignin. Freudenberg 
(1966) formed a quinonemethide, as shown in Figure 2,1, by combining 
radicals that resulted from lignol dehydrogenation. From these experi- 
ments he suggested that quinonemethide was a tentative structural unit 
in lignin. Since the quinonemethide has no opportunity to become 



HgCOH 



HC. 



HC 



HoCOH 



HC 

I! 

HC 




\ 



OMe 



HC 




\ 



OMe 



Fig. 2.1 Quinonemethide 



stabilized by hydrogen migration, it adds on the external electro- 
lytes, particularly hydroxyl compounds and preferably water. Qui- 
nonemethide is a chromophore, is yellow and can be easily recog- 
nized by its intense color absorption extending into the beginning 
of the visible range. Quinonemethide can achieve limited stability 
through polymerization, creating large molecules that can still inter- 
act with polar compounds. 

It is possible to construct a tentative constitutional scheme 
for spruce lignin, which probably is similar to other wood lignins. 
Such a scheme is presented in Figure 2.2. The lignols which origi- 
nate during lignin formation, together with the hydrolysis products, 
reveal different ways in which the C^C_ units are combined. 

Through natural and industrial processes the lignin is separa- 
ted from wood fibers and produces chromophores in aqueous solutions. 
Kirk et al. (1969) prepared lignin by bacterial degradation of 
wood. The lignin was fractionated by molecular gels into three 
separate fractions, all of which would absorb light in the visible 
spectrum. Alder et al. (1966) degraded spruce lignin by acid reflux- 
ing in an organic aqueous solution, and was able to separate through 
fractionation several products that were color producing compoxmds. 

2-3 Characteristics of Color 
Many investigators have attributed the color present in water 
to the natural or induced degradation products of organic matter. 
Saville (1917) concluded through electrophoretic studies that most 
organic color was negatively charged and existed in the colloidal 
size range. Black and Christman (1963a) found that color collected 



10 



H2COH 

— CH 

I 

HC 

I 



o 



HgCOH 



U I 

^^^ — CH 



OMe 



HpCOH 

— CH 

1 
HC 



HoCOH 




OMe 



HC- 



I 

HCO(C6H,o05)nH 




H;>COH 

3 CH 

I 

HCOH 

I 

MeO 




H2COH- 



CO 



MeOjj'^H.OMe 






HC- 
I 




I 

•0 



HoCOH 

I 

HC — 



CO 




HpCOH 

I 

HC — 

I 

HCOH 
I 




OMe 



I 

OH 



1/2 1/2 

HCO HpCOH 

i I 

HC HC — 

HC CO 



OH 



MeO 





HgC 
HC 

I 

HC 






i 



CH 
CH 

I 

CH2 




OMe 



H2COH 
-CH 



HC- 

J 



y — .MeO 
O-OH 



0— 



OMe 



I 

OH 



0-0 /O- 




OH 



I 

HCOH 

i 

CH 



I 

CH- 




OMe 



HgCOH 



CH 

I 

CH 

I 

CO 



COH 

I 

CH 



I 

HC. 

HC^^^i 
HoioH 







\ /"' 



OMe 



Fig, 2:2 Constitution scheme for lignin 



11 



from ten different water samples had similar chemical and physical 
characteristics. They demonstrated by dialysis that most of the color 
present in the ten samples resulted from colloidal suspensions. The 
infrared spectriom for each of the fulvic fractions, the equivalent 
weights of those fractions, and the concentrations of the fulvic 
and humic fractions in each colored sample were similar. Black and 
Christman (1963b) demonstrated that color intensity was pH dependent 
and v/ould increase with increasing pH. They also found by dialysis 
that color existed as a colloid, because only 10% of the original 
sample color could pass a 4.8 micron filter. 

Shapiro (1958) found that organic color was mainly dicarboxylic 
hydroxy aliphatic organic acids of molecular weight 450. He sug- 
gested that if phenols were present they were non-color producing 
organic compounds. He also found that the salts of these acids would 
pass a cellophane membrane, indicating that they were not colloidal. 
Shapiro (1958) demonstrated, by chromatographic comparisons, that 
chemical patterns of color samples taken from different lakes across 
the country were similar. Any differences that existed in these sam- 
ples were due to inorganic constituents of the water. Black (1960) 
suggested that separation techniques used by Shapiro (1958) excluded 
a portion of the color bodies, and that the excluded portion was in 
the colloidal size range. 

Christman and Ghassemi (1966) isolated seven different phenolic 
compounds common to wood and water humics. Their organic analysis 
on wood lignins identified carboxyl and phenol groups as the major 
building units in color molecules. They described these groups as 



12 



large aromatic molecules with hydroxyl, methoxy and carboxylic func- 
tional groups, Christman and Ghassemi (1966] also found that color 
extracted from soil would increase with time of soil contact and 
temperature of the aqueous color medium. Their research showed, as 
had that of Black and Christman(1966) , that color increased with an 
increase in pH. However, this increase was not linear over the 
entire pH range. 

Taylor and Zoltek (1974) , using a kraft effluent treated for . 
color removal by massive CaCOH)^ treatment, found that color increase 
in the waste occurred when the waste was in contact with soil or 
light. The amount of color increase in the soil-contacted samples 
was directly proportional to the organic content of the soil. 
Gjessing and Samdal (1968) studied color fluctuation in a chain of 
four Norwegian lakes and found that color decreased in all of them 
except the last lake, where an impoundment occurred. The last lake 
had a high organic matter content. Gjessing and Samdal (1968) recor- 
ded a direct increase in the color of the impounded lake with time 
of water storage. Their data led to the conclusion that solubilized 
organic matter produced the color increase in the impounded lake, 
and the degree of color increase depended on time of impoundment. 

Packham (1964) separated color from seven different waters into 
the same classes as did Black and Christman (1966). He found, based 
on filtration of the fractions, that the fulvic acid fraction existed 
in the colloidal size range and that the humic acid was in the molecu- 
lar size range. Packham (1968) also revealed that both the fulvic 
and humic fractions consisted of complex mixtures of many different 



13 



organic acids. Gjessing and Lee (1967) fractionated the color pre- 
sent in a natural water by gel filtration and found molecular size 
distributions ranging from greater than 200,000 to as low as 700. 
They found that the molecular size fraction that contained the lar- 
gest concentration of organic carbon did not produce the greatest 
color. 

Midwood and Felbeck (1968) purified a yellow color from organic 
muck and found that ths organic matter producing color was resistant 
to chemical or biological degradation. They found that over 80% of 
the organic carbon was present in the fulvic portion of the color. 
The infrared spectra showed that aromatic carboxylic acids with ali- 
phatic side groups containing phenolic hydroxyl groups were major 
components of the color molecules. Day and Felbeck (1974) obtained 
a yellow water-soluble organic exudate from the domestic waste water 
fungus Aureobasidium pullulans . The exudate contained no himic acid, 
although it was yellow and was very homogeneous relative to fulvic 
acid extracts from soil. Day and Felbeck (1974) demonstrated that 
fungal activity was one source of color in watersheds, and concluded 
that watershed management with respect to excess biological activity 
may help eliminate color problems in watersheds. 

2-4 Coagulation 
The reader is referred to comprehensive literature reviews on 
general coagulation that were published by the American Water Works 
Association (1971) and O'Melia (1972). In this section the emphasis 
will be on coagulation as it refers to color removal. 



14 



A colloidal dispersion is electrically neutral, so that the 
charges on the colloidal surface must be counterbalanced by the 
charges on the liquid immediately adjacent to the colloidal parti- 
cle. As a result, an electrical double layer exists at every solid- 
liquid interface. These charged ions are attracted to the colloidal 
surface electrostatically and repelled due to diffusion. The Verweey- 
Overbeck model as described by Osipow (1972) stated that the London- 
Van der Waals forces were the forces of attraction for colloids in 
suspension. The forces of repulsion resulted from the electrical 
repulsion of the separate colloidal double layers. Osipow (1972) 
demonstrated that this model was. further developed and modified by 
Guoy, Chapman, Stern and Helmholtz. The essence of the final model 
was that colloidal suspension would be destabilized if the electrical 
repulsive forces were reduced such that the London -Van der Waals 
forces would dominate, causing coagulation and sedimentation. This 
concept was supported in some systems by the Schulze-Hardy rule, 
which states that the critical coagulation concentration of mono-, 
di- and trivalent ions to coagulate sols of the opposite charge are 
in the ratio of 100:1.6:0.13. Matijevic et al. (1964a) developed 
a stabilization-destabilization model for AgBr and Agl suspensions 
based on neutralization of the electrical double layer with counter 
ions gained from the hydrolysis of A1(N02)2. Matijevic et al. 
(1964b) attributed the destabilization of the sols to the Al spe- 
cies on the basis of charge reversal in the coagulation reaction. 
However, there was a stabilization of the sol which was followed 
by another sol coagulation. Matijevic et al . (1964b) contributed 



- 



15 



the final destabilization to Al (OH) precipitation. They presented 
no explanation for the restabilized sol prior to Al (OH) precipita- 
tion, since the sol charge remained positive after the first coagu- 
lation. 

LaMer (1967) developed a bridging theory which provided an 
acceptable qualitative model for describing the destabilization of 
colloids with polymers. The main points of the bridging theory 
were: 1) the polymer must contain chemical groups that would interact 
with the colloidal surface; 2) that when this happens only a part of 
the colloidal surface was covered, and the remainder of the polymer 
would serve as a bridge upon attachment to another colloid; 3) if no 
other colloid was available for attachment, or the polymer concentra- 
tion was too great, the polymer would attach itself to the colloid 
and restabilize the suspension; 4) intense agitation would sever the 
polymer bonds to the colloidal surface and possibly restabilize the 
suspension; 5) the amount of colloidal surface area present was 
directly proportional to the amount of polymer required for coagula- 
tion. The bridging theory explained how chemical interactions 
between an anionic polymer and negative colloid would produce 
coagulation. 

Packham (1968) studied coagulation of eight different clays by 
aluminum hydrolysis and found the coagulant dose continually decreased 
with increasing concentration. Solubilized calcium and magnesium 
assisted in lowering the coagulant concentration necessary to desta- 
bilize clay suspensions. Packham (1968) demonstrated that the hydro- 
lysis products of cilusa were important to clay destabilization by 



" Tft'fllW ' tUd" tfW I 



16 



zeta potential measurements o£ clay suspensions dosed with and with- 
out alum. Although the alum floe had the same zeta potential for 
optimum destabilization as did the clay suspension, the zeta potential 
was not zero. Apparantly electrostatic forces were not controlling 
destabilization. Schott (1968) studied the deflocculation of water 
sorping clays by anionic and nonionic surfactants. He found that maxi- 
mum deflocculation was produced when the surfaces of the clay lattices 
were completely covered with the nonionic surfactants. 

The American Chemistry Society (1968) published data for aluminum 
hydrolysis in colloidal suspensions showing that the polynuclear spe- 
cies of aluminum were important destabilization factors. A colloidal 
suspension was destabilized before any floe was formed using alum as 
the coagulant. They suggested the forces of adsorption between the 
colloids and the hydrolysis products were responsible for destabili- 
zation, because the hydrplyzed species were hydrophobic and were more 
likely to accumulate at the solid-liquid interface. Another factor 
leading to colloidal destabilization was that the hydrolysis products 
had more than one OH" ion that could sorp at the interface. Their 
data indicated that as the colloidal surface area concentration 
increased, an increasing coagulant dose was required to destabilize 
the colloidal suspension. 

Langelier and Ludwig (1949) experimented with calciiom and alum 
flocculation of four different turbid waters varying in exchange capa- 
city. They concluded that the mechanism of colloidal destabilization 
was controlled by the exchange capacity of the colloids. Michaels 
(1954), studying the degree of polymer hydrolysis that best promoted 
coagulation, found that a small amount of hydrolysis was best suited 



17 



for destabilization. He suggested that the destabilization mechanism 
was a two step process: 1) polymer sorption onto the colloidal sur- 
face and 2) interparticle bridging following polymer sorption to 
destabilize the colloidal suspension. 

Black et al . (1965) evaluated coagulation by anionic polymers 
and demonstrated destabilization followed by restabilization with 
excess polymer concentration. Since both the polymer and the colloid 
were negatively charged, the destabilization was not attributed to 
coulombic forces, but to the build-up of interparticle bridges 
through other than electrostatic mechanisms. They also found that a 
higher velocity gradient for a shorter time period was more effective 
in destabilization than a lower velocity gradient for a longer time 
period. Ragunathan et al . (1973) treated turbid waters with alum 
and concluded, from zeta potential measurements, that the hydrolysis 
products of alum were controlling destabilization by sorption mechan- 
isms. 

Posselt et al. (1968a) examined metal sorption onto a Mn02 
anionic sol and found neutral and anionic species did not sorp, a fact 

supporting an electrostatic mechanism for destabilization. Posselt 

++ 
et al. (1968b) studied Ca sorption onto a negative Mn02 sol and 

found Ca sorption onto the Mn02 colloidal surface approached a limi- 
ting value. The limiting Ca sorption indicated a Langmuir monolayer 
was probably occurring on the Mn02 surface. They restabilized the sus- 
pension with more polymer addition, but did not achieve restabilization 
with increased metal ion concentrations. However, the increased cal- 
cium concentration did broaden the optimum range for coagulation. 



18 



They suggested that a choice of coagulant aid would be based on the 
potential determining ions of the sol. Robinson et al. (1974) repor- 
ted that larger increases in the turbidity of a river water increased 
treatment costs and large quantities of alum were required to produce 
potable water. Nonionic and cationic polyelectrolytes were found to 
be more effective than alum, suggesting that for this water an electro- 
static mechanism was not controlling destabilization. Aluminum hydro- 
lysis was probably removing turbidity by enmeshment in a sweep floe. 

LaMer (1967) defined coagulation as a kinetic process going from 
a quasi-stable to a more stable phase, and flocculation as the 
bridging of already coagulated particles that entered into hindered 
settling. As an example, he cited hydroxyl groups on flat clay sur- 
faces bonding with hydroxyl radicals of polymers, which allowed metal 
ions to form insoluble phosphates. LaMer (1967) suggested turbidity, 
subsidence rate and floe filtration as methods of evaluating desta- 
bilization. He also suggested that a negative polymer would best 
destabilize a negative colloid, because many sites were produced by 
polymer hydrolysis for bridging. 

Birkner and Morgan (1968) measured particle size distribution 
during coagulation and found stronger floe was produced as floe 
diameter increased. They demonstrated the rate controlling step 
was particle agglomeration after coagulation, and that intense agi- 
tation was responsible for limited polymer sorption. Dollimore and 
Horridge (1972) investigated flocculation of China clay using poly- 
aerylamides. They found that the maximum clarity was not coincident 
with the maximum filtration rate as measured by the Kozeny-Carmen ' 



■^l»1 if »^'R M 4^f^1r-m' O ^V9'tmm4 ^ -maire^^—,~Urjii:^..^:iL^.-^ 



19 



equation. They concluded that the effective length of the floc- 
supernatant interface was the controlling flocculation parameter. 
Hahn and Stumm (1968) , studying the kinetics of alum hydrolysis 
for SiO sols, determined that there were three steps in the coagu- 
lation process: 1) forming polynuclear hydrolysis products; 2) the 
rate of surface coverage or adsorption of the polymer on the colloi- 
dal surface; and 3) the rate of particle transport. The rate limi- 
ting step for Si02 coagulation was shown to be the rate of particle 
transport. The rate of coagulation was sho\m to be a function of 
the collison rate and the collison efficiency. 

Tenney and Stumm (1965) demonstrated that hydrolyzing metal ions 
and organic polymers could be used to successfully coagulate bacteria. 
A linear relationship was found between the optimum concentration of 
the polyacrylamide polymer and the bacterial concentration. They 
also found that phosphates were removed with Al in a chemical 
reaction, and the optimum pH for the reaction was the same as for 
optimum bacterial flocculation. 

Stumm and Lee (1961) found that the rate of oxidation of ferrous 
iron was directly proportional to pH. They found an increase of one 
pH unit near neutral pH resulted in a 100 fold increase in oxidation 
rate. Schenk and Weber (1968) also determined that the rate of oxi- 
dation of ferrous iron increased with increasing pH. They found 
that silica retarded the hydrolysis of Fe . The hydrolysis was 
not represented by a first order reaction, but approached linearity 
with time. They suggested that the solubility relationship may have 
been altered by complexes formed between silica and iron, and that 



■jj^i i iw . i i iii*iiiia 



20 



these complexes may have been the mechanism by which activated silica 
functioned as a coagulant aid. Mohtaoi and Rao (1973) investigated 
the effects of temperature on aqueous suspensions and concluded that 
temperature had no perceivable effect on the zeta potential of the 
sols, or the alum hydrolysis products mixed with cationic, anionic and 
nonionic polymers. Charge neutralization was determined to be impor- 
tant in destabilizing a colloidal suspension. The neutralization had 
to be achieved before flocculation occurred. The optimum pH for alum 
coagulation was found to vary with temperature. However, coagulation 
with cationic polyelectrolytes was found to be temperature independent 
of the flocculation rate, optimum pH and coagulant dose. 

Stumm (1967) demonstrated that metals acted as Lewis acids and 
had a tendency to stabilize pH. He described metal ion hydrolysis 
as a function of pH and metal ion concentration. St-umm (1967) stated 
that multivalent hydrous oxides were amphoteric and that H'*' and OH" 
were primarily the potential determining ions for such hydrous oxide 
precipitates. He also stated that metal ions precipitated in the 
presence of coordinating anions usually as nonstoichiometric mixed 
precipitates. Stumm et al. (1967) formed polysilicates and classi- 
fied them into three separate areas: 1) insoluble, 2) stable poly- 
mers and 3) the mononuclear wall. They concluded from potential 
measurements that the interaction between the anionic polymeric 
phase and the negative sol was due to specific sorption and would 
overcome electrostatic repulsion. They found optimal destabiliza- 
tion occurred when a fraction of the colloidal surface area was 
covered and suggested that the mechanism of destabilization for 
activated silica was the same as for polyelectrolytes. 



21 



Sturam (1967) published a hydrolysis model for colloidal desta- 
bilization that accounted for bridging and electrostatic effects. 
He postulated that a fraction of the total colloidal surface area 
must be covered to produce coagulation. He expressed the model 
mathematically using a Langmiur isotherm by equating the amount of 
coagulant necessary to produce a certain fractional coverage to the 
sum of the residual and sorbed coagulant. The fractional surface 
coverage necessary to destabilize colloidal sols could only be 
gained from the residual coagulant or the sorbed coagulant. He 
showed from his model that the required coagulant dosage to produce 
destabilization could be independent of surface concentration or 
linearly dependent on surface concentration. In the Stumm model 
metals first destabilized colloids due to sorption of the hydrolyzed 
cationic coagulants and restabilized the colloids due to extensive 
sorption of the hydrolyzed metal coagulants. Finally a precipita- 
tion of the metal occurred that destabilized colloids. If the coagu- 
lant became attached to the colloidal surface, the coagualnt dose 
decreased with increasing colloid concentration. In the precipi- 
tation zone the coagulant enmeshed the colloids in a sweep floe. 
If this occurred, the coagulant dose was not a function of the 
colloidal surface area. If the colloidal concentration was high, 
the amount of coagulant dosed could be such that initial destabili- 
zation by sorption and final destabilization by precipitation would 
be indistinguishable. 

The Stumm model for a large colloidal surface area predicted 
a large nonstoichiometric coagulant dose that could be reduced if 



22 



buffering were removed. A system with a medium colloidal surface 
area requir^ed a stoichiometric coagulant dose, and if buffering were 
present, the zone of coagulation was reduced. If low colloidal sur- 
face area were present, a large nonstoichiometric dose would be 
required to coagulate by precipitation. Stoichiometry could be 
achieved through alkalinity additions. 

Kawamura (1973) reported that Ca (OH) 2 additions should be made 
after or during alum coagulation for optimum turbidity and color 
removal. Jeffcoat and Singley (1975) found that Ca(OH) addition 
prior to alum coagulation increased turbidity removal and recommended 
doing so for optimum coagulation results. 

Hannah et al. (1967) measured alum floe size variations with 
kaolin, polymer and pol>'phosphate additions. They found kaolin and 
polymers increased floe size. Polyphosphates hindered floe forma- 
tion. They recommended that the polyphosphates should be added last 
in the coagulation process. Hannah. et al. (1967) demonstrated that 
the order of chemical addition affected the coagulation process. 

Olson and Twardrowski (1975) studied the products formed by coa- 
gulating high alkalinity waters with ferric hydrolysis and concluded 
FeC03Cs) may be precipitated instead of Fe(0H)2(s). Guilledge and 
0' Conner (1973) found arsenic was removed by both alum and ferric 
chloride hydrolysis. They concluded that adsorption was the removal 
mechanism. Their results indicated that arsenic was removed better 
by alum than iron coagulation, the removal was pH dependent and could 
possibly be the result of a chemical reaction. Stumm and Morgan 
(1962) found, when doing alkalimetric titrations, that the amount of 



23 



base required to titrate the aluminiim mixture was not increased stoi- 
chiometrically in the presence of a pyrophosphoric acid. Their data 
suggested that phosphate removal by alum coagulation resulted from a 
chemical reaction. Cornwell (1975), studying alum recovery through 
liquid-liquid extraction, suggested that phosphate was removed by a 
chemical reaction producing an aluminum hydroxy phosphate. 

2-5 Color Removal by Coagulation 
Black et al, (1963) demonstrated that C(5lor present in six 
different natural waters was removed stoichiometrically by ferric 
sulfate coagulation. A graph of raw water color verses required 
coagulant dose was constructed, and the optimum conditions for color 
removal did not produce a floe that had zero zeta potential. Singley 
et al . (1967) also found that, to obtain maximum color removal, alka- 
linity had to be added before coagulation. Ferric sulfate proved to 
be a better color removing coagulant than alum for the six natural 
waters tested. . 

Packham (1965) studied coagulation of organic color that was iso- 
lated from river water. He separated the color into humic and fulvic 
fractions. The mechanisms of alum and ferric coagulation were found 
to be similar, because stoichiometric amounts of these coagulants 
were required to remove different concentrations of humic and fulvic 
acids. Packham (1965) proposed from his data that humic acid was 
entering into a chemical reaction with aluminum. He determined the 
empirical formula for such a reaction was Al(OH) • R. He found that 
the fulvic portions were more complex than the humic acid portions 



24 



and found little evidence o£ color enmeshment in the Al (OH) _ £loc. 
Packham (1965) did achieve an optimum pH for color removal . Jobin and 
Ghosh (1972) studied the oxidation of ferrous iron. They found that 
the addition of humic acid complexed the ferrous iron and retarded 
the oxidation reaction. Schnitzer (1971) found, at pH 2.5, that 
insoluble fulvic acid precipitates were formed with aluminum only 
when more than one metal ion was added for each carboxylic group 
present. Mangravite et al . (1975) conducted experiments on humic 
acid removal by alum coagulation. They demonstrated that insoluble 
aluminum humic precipitation formed at a pH lower than did pure 
Al(OH),(s) precipitates. They suggested that color was removed from 
solution in alum coagulation by a chemical reaction. 

Narkis and Rebhum (1975) concluded that the salts of humic and. 
fulvic acids acted as anionic polyelectrolytes that reacted chemically 
with the cationic flocculant, the carboxylate and the phenolate groups. 
The reaction products formed a colloidal precipitate that could be 
removed by settling after flocculation. The first step in humic and 
fulvic acid coagulation was suggested to be a chemical reaction before 
flocculation by cationic polyelectrolyte addition occurred. 

Luner and Dence (1970) determined that the color bodies present 
in a kraft waste were mostly aromatic and quinoid nuclei with carboxyl 
or ethylenic groups. The color bodies removed in Ca(0H)2 treatment 
were carboxylic, phenolic or enolic groups that had precipitated in 
a chemical reaction with calcium. They found that both the precipi- 
tated fractions and the nonprecipitated fractions of kraft waste were 
acidic, but that the nonprecipitated fractions were more acidic and 



25 



had a lower average molecular weight than the precipitated fractions. 
Luner et al, (1970) found, with massive lime treatment, that enolic 
groups reacted chemically with calcium to produce insoluble precipi- 
tates. 

2-6 Magnesium Coagulation 

Stumm (1968) demonstrated that metal cations such as magnesium, 
aluminum or calcium could function effectively as coagulants. Magne- 
sium hydrolyzes significantly at pH values encountered in lime soften- 
ing and produces a voluminous floe which hinders solid handling opera- 
tions. Eidsness and Black (1957) reduced the volume of sludge pro- 
duced in water r-oftening operations at Dayton, Ohio and Gainesville, 
Florida by bubbling CO2 into the sludge to dissolve Mg(OH)T Sixty 
per cent of the Mg(OH) was solubilized, but no attempt at optimi- 
zation of magnesium recovery was made. The sludge settled more 
readily after carbonation. Eidsness and Black (1957) concluded that 
because Mg(OH)„ existed as a gelatinous coordination complex it 
could accept a proton more readily than the lyophobic crystals of 
CaCOj. This enabled the sludge volume to be reduced. Black (1971) 
suggested that not all the Mg(0H)2 could be removed from lime soften- 
ing sludge because of MgC0_*3H20 precipitation in the carbonation 
tank. He proposed the use of a floatation process to remove clay 
from the lime softening sludge before recalcination. 

Thompson et al . (1972a) treated potable water samples in the 
laboratory using magnesium carbonate as the coagulant. They were 
able to develop an equation relating coagulant dosage to raw water 



26 



color and turbidity when Ca(0H)2 was used to control' pH. Thompson 
et al. (1972a) proposed a potable water treatment process in which 
magnesium was used as the primary coagulant, and Ca(0H)2 was used 
to control pH. The magnesium was recovered from the sludge by car- 
bonation, and the Ca(0H)2 was recovered from the remaining CaCO^ by 
recalcination. They proposed that sludge handling problems associa- 
ted with conventional coagulation plants would be greatly reduced 
utilizing the magnesium carbonate process. Thompson et al. (1972b) 
compared conventional coagulation systems with the proposed magne- 
slum carbonate system. They also demonstrated that as turbidity and 
color were reduced, the zeta potential of the residual floe was 
increased. 

Dubose et al . (1973) successfully extended the magnesium car- 
bonate process to treatment of domestic sewage in a pilot plant at 
Gainesville, Florida. .He found that magnesium coagulation reduced 
the total phosphorous to less than 0.1 mg/1 P, and significantly 
reduced the suspended solids, color and oxygen demand of the domestic 
wastewater. 

Black (1974), in pilot plant studies at Melbourne, Florida, 
found evidence that the color was released from magnesium sludge 
upon carbonation. This color release was found to stabilize with 
time, which implied that color release in magnesium recovery may 
not be a problem. Studies by Taflin et al. (1975), using CO2 gas 
to redissolve magnesium solids in a lime softening sludge, were 
discontinued due to a high color return with the recovered magnesium. 
The potable water produced by using the recovered magnesium as a 



27 



coagulant was too colored to be acceptable. 

Predali and Cases (1973) investigated zeta potential o£ magne- 
siiim carbonates in electrolytes and found OH' and H"*" to be the poten- 
tial determining ions for MgC0H)2. The Mg(0H)2(s) colloids had a 
zero zeta potential at the same pH for varying ionic strength aqueous 
solutions. They concluded from kinetic considerations that MgOH 
must have been the source of the positive charge on the Mg(0H)2Cs) 
colloid. Zoganathan and Maier (1975) found that sand and kaolinite 
colloids in a solution of 0.005 M MgCl2 had a positive zeta potential 
for a pH of 10.3 or greater. They attributed the positive zeta poten- 
tial to the increasing percentage of MgOH"^ relative to the total spe- 
cies of soluble magnesium. 

2-7 Color in Pulp Mill Effluents 
Fitzgerald, Clemens and Riley (1970) demonstrated that while 
polymers could destabilize colloids in pulp waste, they would not 
neutralize the electrical double layer. Zettlemeyer, Micale and 
Dole (1968) studied sludges from pulp mills and found that floccula- 
tion kinetics varied with pH for organic carbohydrate base sludge, 
but not with an inorganic primary sludge from a newsprint mill. 
Their data indicated that most of the colloidal bound water was 
interstital and was not chemically held due to the solid-liquid 
interface. The National Council for Air and Stream Improvement 
(1971) studied surface properties of hydrogels resulting from treat- 
ment of pulp mill waste and found anionic polymers destabilized 
negative colloids. They suggested that the polymer sorption onto 
the negative surface was nonstoichiometric. 



28 



Davis (1972), using CaCOH)^ coagulation 'at Riceboro, Georgia 
to remove color from a kraft waste, found calcium solubility decrea- 
sed as the sodium concentration from the digestion operation was 
increased. Davis demonstrated that organic carbon, color and calcium 
concentrations after Ca(0H)2 treatment were related to the initial 
sodium concentration of the waste. Berger (1964) found that a large 
Ca(OH) dose (15,000-25,000 mg/1) produced a very settleable floe 
that removed 90% of original color from caustic bleach effluent. The 
Domitar Limited Research Center (1974) reported that Ca(0H)2 coagula- 
tion was not effective for removing color from sulfite liquors. The 
Interstate Paper Corporation in Riceboro, Georgia used a smaller 
chemical dose of Ca(0H)2 (1500-2500 mg/1) to remove in excess of 90% 
of the initial color in a kraft waste. However, the lime dose did 
exceed the solubility product of Ca(0H)2 and formed a precipitate. 
Othof and Eckenfelder (1974) studied color removal from three kraft 
mill effluents by separate coagulation with Ca(0H)2, ferric sulfate 
and alum. They suggested that ferric sulfate was the better coagu- 
lant because of lower coagulant dose and less voluminous sludge vol- 
umes. Gould (1973) reported that the effluent from the caustic 
extract stage of a kraft bleach plant, when treated with Ca(0H)2, 
would form a metal organo precipitate that removed 90% of the initial 
color. Approximately 80% of the Ca(0H)2 was recovered in the sludge. 
Spruill (1975) found Ca(OH) treatment was very effective for reducing 
color in kraft wastes, but was ineffective for removing color from 
sulfite waste. Leszczynski (1972) concluded that of the many pro- 
cesses proposed for color removal from kraft wastes, only Ca(0H)2 



29 



precipitation was feasible. Kabeya et al. (1972)' found the rate of 
absorption of kraft mill lignins on activated carbon to be very low. 
Katoh and Kimura (1972) found fly ash to be almost as effective as 
activated carbon in sorping lignin from kraft mill effluents. 

The National Council for Air and Stream Improvement (1974) 
studied the mechanism of color removal on activated carbon and found 
most color bodies existed in the high molecular weight (15,000) 
range. TOC and color were not removed in equal proportions. They 
concluded that color removal by activated carbon was not a chemical 
process, but was due to soi-ption. Swanson et al. (1973) did a 
detailed study on Ca(6H) treatment of kraft waste and found 86% 
color reduction, 57% TOC reduction and 17% sugar reduction. There 
was no removal of material with molecular weights less than 400. 
Material with molecular weights greater than 5000 was completely 
removed, and partial removal was observed for material with molecu- 
lar weights ranging from 400 to 5000. Swanson (1973) suggested 
color bodies were aromatic groups that carried a negative charge. 
Tejera and Davis (1970) used alum, AlCl^ and FeCl^ as coagulants in 
color removal studies on caustic extraction waste and chlorinated 
waste. They determined both FeCl and AlCl were capable of removing 
96% of the color from a kraft mill caustic extraction waste, but both 
coagulants were hampered in the removal of color from the chlorinated 
waste. 

Collins et al . (1969) separately concentrated chlorinated and 
alkaline extraction bleach effluents from a sulfite and a kraft pro- 
cess by reverse osmosis. Lignosulfonic acids with molecular weights 
in excess of 10,000 were found in the sulfite waste liquor. Jensen 



30 



et al. (1964) fractionated spent sulfite waste liquors by gel filtra- 
tion and ion exclusion and found six different components. Saccah- 
rides and weak organic acids at pH 4 were present in the lower mole- 
cular weight range. Fractions above a molecular weight of 40,000 
were aromatic lignosulfonic acids and were responsible for most of 
the color in the waste. 

Smith and Christman (1969) 'treated kraft and sulfite waste with 
AlCl^ and FeCl_ coagulants and found either coagulant would remove 
90% of the initial color in the kraft waste. Treatment of the sul- 
fite waste by FeCl^ reduced the organic carbon 50%, but increased 
the color of the sulfite waste. Alum reduced the color of the sul- 
fite waste 67%. Smith and Christman (1969) proposed that the kraft 
waste had sulhydryl groups on lignin chains and that these groups 
formed insoluble sulfides during coagulation. The sulfite waste 
had sulfonate groups in the lignin chain which acted as strong acids 
and formed hydrolysis products. The mechanism for color removal in 
the kraft waste was suggested to be a chemical reaction, whereas the 
mechanism in the sulfite waste was suggested to be sorption on Al (OH) ^ 
surfaces . 

Rapson et al . (1971) used seawater as a source of soluble mag- 
nesium along with Ca(OH) to remove color from a kraft waste. Increa- 
sed color removal was accompanied by the formation of a floe with a 
larger surface area than the original Ca(OH) floe. A 20% seawater 
mixture did not remove any more color from the kraft waste than did 
a 10% seawater mixture. Less color removal was observed when a card- 
board effluent was treated, which indicated that different mechanisms 



31 



might have been responsible for color removal" for different kraft 
effluents. The Canadian Pollution Abatement Research Program (1974) 
used Ca(OH)' and MgCl^ to treat sulfite waste for color removal. 
They obtained an 86% reduction of NSSC waste using Ca(0H)2 and MgCl2 
and a 65% reduction of color using MgCl2 without Ca(0H)2. They were 
able to remove 86% of the color from a bleach-kraft, unbleached 
kraft, combined bio-kraft, NSSC-NH3 base and a bio-NSSC waste. They 
did not optimize pH or coagulant dose in the coagulation process. 



CHAPTER 3 



LABORATORY PROCEDURES 



3-1 Feed Solutions 



3-1.1 Synthetic Waste Solutions 

All wastes were made by diluting a concentrated color source 
with tap water to the desired color concentration. The color source 
was a stored semichemical neutral sulfite liquor which was taken 
from a NSSC plant digestor after the cooking operation had been com- 
pleted. This liquor contained the dissolved constituents of the 
wood. It was referred to as "sulfite waste liquor," which can be 
the major source of color in the waste stream of a neutral sulfite 
semichemical pulp plant. The sulfite waste liquor was obtained from 
plants located in Harriman, Tennessee and Hartsville, South Carolina, 
owned by Mead Corporation and Sunoco Products Company respectively. 
The Mead Corporation supplied soldium base spent sulfite waste liquors 
and ammonium base spent sulfite waste liquors that were used as a 
waste source. The Sunoco Products Company supplied a sodium base 
spent sulfite waste liquor which was also used as a waste source. 

There are different processes and many different types of hard- 
wood trees used in neutral sulfite semichemical pulping. Because of 
this, it was decided at the beginning of this research to determine 

32 



33 



if color could be removed from different semichemical neutral sul- 
fite wastes by magnesium coagulation, but only to use the sodium 
base waste from the Hartsville, South Carolina plant to study mech- 
anism of color removal by magnesium coagulation. 

The color of the stored NSSC spent liquor varied from 250,000 
to 500,000 Pt-Co color units. Consequently, to achieve a working 
color of 5000 Pt-Co color units, a dilution ratio of 50/1 to 100/1 
was required. 

3-1.2 Coagulation Chemicals 

Magnesium sulfate, MgS0.-7H„0, was used as the source of magne- 
sium ions for the color removal process. A stock magnesium solution 

++ 
of 50 mg/ml as Mg was made in order to minimize the volume of 

coagulant feed dosed in the process. This was achieved by dissol- 
ving 532.6 grams of MgSO^ "7^120 in a liter of distilled-deionized 
water. 

Calcium hydroxide and sodium hydroxide were used for pH adjust- 
ment during the coagulation reaction. Calcium hydroxide was slurried 
in a small beaker before it was used, whereas sodium hydroxide was 
added from previously prepared 10 N and 1 N solutions. When neces- 
sary, sulfuric acid and hydrochloric acid were used to adjust the 
pH downward. 

3-1.3 Polymers 

Cationic, anionic and nonionic polymers were prepared from 
commercial liquids and powders supplied by American Cyanamid Company. 
The polymers were made from polyacrylamide and amine bases. Stock 



34 



solutions of 2000 to 3000 mg/1 were prepared from the solid based 
polymers by choosing a weighed amount and dissolving it in an aqueous 
solution by magnetically stirring it overnight. The liquid based 
polymers at the same concentrations required only one hour of stir- 
ring for stock preparations. For the colloidal, acids, an activator 
supplied by American Cyanamid CN-478) was required for stock prepara- 
tion. 

3-2 Analytical Equipment and Techniques 

3-2.1 Total Carbon Measurements 

Total carbon measurements were determined on a Beckman Model 
915 Total Carbon Analyzer in conjunction with a Beckman Model 865 
Nondispersive Infrared Analyzer. A three microliter sample was used 
for analysis. The readout was registered on a to 100 scale and 
was compared to a standard curve. The carbon standards were pre- 
pared from potassium biphthalate for organic carbon, and sodium car- 
bonate or sodium bicarbonate for inorganic carbon. 

3-2.2 Color Measurement 



All color measurements were determined according to NCASI Tech- 
nical Bulletin 253. This procedure requires all samples for color 
measurement to be filtered through a 0.80 micron Millipore filter. 
The pH of the sample was then regulated to 7.6 before the amount of 
absorbance at a wavelength of 465 millimicrons was recorded. The 
sample color was then calculated by locating the sample absorbance 



35 



on a standard curve relating color to absorbance. If the sample had 
too great a color to be directly measured, the sample was diluted 
after filtration. 

A standard curve was prepared by dissolving 1.246 grams potas- 
sium chloroplatinate, K2PtClg (equivalent to 0.500 g metallic plati- 
num) and one gram crystallized cobaltous chloride, C0CI2 '61120 (equi- 
valent to 0.25 grams metallic cobalt) in distilled water with 100 ml 
concentrated HCl. This solution was diluted to 1 liter with dis- 
tilled water. This stock solution was defingd as having a standard 
color of 500 Pt-Co units. A standard curve was prepared and is shown 
in Figure 3.1. This curve fits the equation: 

Color = (2183.4) (absorbance) -^ 4.4 ' (3-1) 

3-2,3 Incineration 

Sludge incineration was determined in a Thermodyne furnace. 
Model F-A1730. Sludge samples were dried at 103°C and filtered 
through a Buchner funnel on a Whatman no. 40 ashless filter before 
incineration in the furnace. Incineration temperatures were varied 
from 180°C to 850°C. Times of incineration were varied from 15 
minutes to 120 minutes. 

3-2.4 Jar Tests 

Jar tests were performed on a Florida Jar Test Machine. Chemi- 
cals were dosed simultaneously to four 1 liter beakers. Rapid mixing 
took place at 100 rpm for three minutes. The Flo.rida Jar Tester was 
capable of 145 rpm, but due to the heavy floe formed when treating 



36 




Color 



Fig. 3.1 Standard Pt-Co color curve 



37 



the highly colored waste, a stirring rate o£ 100 rpm was the maxi- 
mum that could be attained. Beyond 100 rpm, the magnetic couple 
between the jar stirrers and the machine was broken by the stirrer 
over-turning. Slow mix took place at 35 rpm for 15 minutes. The pH 
was adjusted through both the slow mix and the rapid mix cycles to 
maintain a constant pH during the coagulation reaction. Floe was 
allowed to settle for 30 minutes before samples were taken for analy- 
sis. If the coagulating mixture had not developed a clear superna- 
tant, a sample was taken and filtered through a no. 40 Whatman fil- 
ter in order to simplify the required filtering step through the 
0.80 micron Millipore filter before color measurement. 

The G levels of the rapid mix and the slow mix cycles were 110 
sec"-^ and 30 sec"-*" respectively. The mixing level in the floccula- 
tion stage had a Gt value of 27,000, which was approximately the low 
end of the range specified in Waste Treatment Plant Design (1971) . 

3-2.5 Metal Analysis 

Metal analyses were determined on a Varian Techtron Model 1200 
Atomic Absorption Spectrophotometer. Magnesium measurements were 
made at a wavelength of 202.5 nanometers. Calcium measurements were 
made at a wavelength of 422.7 nanometers. A.ll samples were filtered 
through a 0.80 micron Millipore filter and treated with 1 ml of 17% 
lanthium-HCl solution per 10 ml of sample before calcixim and magne- 
sium values were measured. 

3-2.6 Mobility Measurements 

All mobility measurements were made with a Zeta Meter. The Zeta 



38 



Meter was .used in conjunction with a Riddick cell and a stereoscopic 
microscope. The multiscale 15X ocular micrometer was used in the 
Zeta Meter for mobility measurements. . A platinum-iridium cathode 
coupled with a molybdenum anode were used in the Riddick cell. 

3-2.7 pH Measurements , 

All pH measurements were made on a Corning Model 12 expanded 
scale pH meter. A Corning silver-silver chloride reference elec- 
trode in conjunction with a glass electrode were used for all pH 
measurements. 

3-2.8 Settling Tests _. 

All settling tests were conducted in a standard 1000 ml gradua- 
ted cylinder. One liter of waste was coagulated in a jar and imme- . 
diately transferred to a graduated cylinder where the height of the 
sludge-supernatant interface was recorded. The following formula 
was used to calculate the Sludge Volume Index, SVI: 

SVI = ml settled sludge x 1,000 
mg/1 suspended solids 

3-2.9 Solids Analysis 

All suspended solids analyses were determined on samples that 
were filtered through a no. 40 Whatman filter and dried at 103 C for 
one hour. Nonvolatile and volatile solids were by filtering the 
samples through a no. 40 ashless Whatman filter, drying at 103 C for 
one hour, and recording the weight. The sample was then ignited at 
550°C for 60 minutes, after which it was weighed to determine 



39 



nonvolatile solids. Samples were weighed immediately after drying 
at i03°C, but were cooled for one hour in a dessicator after igni- 
ting at 550°C before .weighing. 

3-2.10 Titration Curves 

0.01 and 0.1 N H2SO4 and NaOH were used for determining the 
acid-base strength of the samples. The volume of waste titrated 
varied from 50 to 200 ml. One minute was allowed for pH stabili- 
zation each time the titrant was added to the ^sample. A Teflon 
covered magnetic bar in conjunction with a magnetic stirrer were 
used to mix the solution during titration. 

3-3 Experimentation 

3-3.1 Coagulation Experiments 

The color removal experiments began by mixing the waste to the 
desired color concentration and measuring the color as previously 
described. The next step was regulation of the waste solution pH 
with CaCOH)- or NaOH. The coagulant was then added and the reaction 
pH was adjusted. Samples for analysis were taken after 30 minutes 
of settling. Organic carbon measurements and color values were 
determined immediately after coagulation. The samples to be analyzed 
for metal concentration were acidified immediately following coagu- 
lation. 

3-3.2 Coagulant Recovery 

Several methods were used to recover the magnesium coagulant. 



40 



Pollowing coagulation the resulting sludge was filtered through a 

Buchner funnel and was then dried at 103°C for one hour. The dry 

sludge was ignited at 550°C, and the resulting nonvolatile solids 

were placed in contact with a 10% CO2 gaseous stream or stabilized 

with H SO to recover the oxidized magnesium. Two 40 liter volumes 
2 4 

of waste were treated in order to produce a large quantity of sludge. 
These wet solids were then heated until they achieved a constant weight 
at 103°C. The dried solids were ignited at 550°C to remove the coagu- 
lated color. 

3-3,3 Coagulant Recycle 

The magnesium was recycled to determine the effectiveness of 
reusing the same magnesium as the primary coagulation in the color 
removal process. The method of recycling the magnesium consisted of 
incinerating the sludge produced in the coagulation reaction at 
550°C. The resulting nonvolatile solids were carbonated for 45 
minutes with a 10% CO2-90% air gaseous mixture. The recovered magne- 
sium was recycled with and without the nonvolatile solids that were 
not dissolved during carbonation. The two different techniques of 
recycling the magnesium determined the effectiveness of the remaining 
nonvolatile solids in the color removal process. 



CHAPTER 4 



RESULTS 



4-1 Determination of Coagulation pH and Coagulant Dose 
Development o£ the color removal using magnesium coagulation 
required that the pH control agent, coagulation pH and coagulant 
dose be determined. Thu chemicals selected for pH control were 
Ca(OH) and NaOH because they were inexpensive and commercially 
available. The coagulation pH and- coagulant dose were defined as 
the minimum pH and dose that resulted in a 90% reduction of the ini- 
tial color. .; . 

A three step technique was used to determine the coagulation 
pH and coagulant dose for the color removal process. First the coagu- 
lation pH was found by determining the reaction pH where maximum color 
removal occurred for a constant magnesium dose. Tlie second step was 
to determine the magnesium dose that removed 90% of the color at the 
coagulation pH. Finally the stability of the coagulation pH was 
verified by repeating the first step for a magnesium dose other than 
the coagulation dose. If the coagulation pH did not shift, then the 
coagulation pH and coagulant dose were acceptable. If the shift in the 
coagulation pH occurred, then the coagulation pH had to be determined 
as a function of both the coagulant dose and the initial color. 

4-1. .1 Coagulation pH 

Figures 4.1 and 4.2 show the curves from which the coagulation 
pH can be determined using Ca(0H)2 or NaOH for a NSSC waste with an 

41 



42 



4000 


- . 


Color = 2500 




' 1 


I 0= Mg"^= 150 mg/l- Ca(0H)2 


3000 




\ •= Mg* = 750 mg/l - NaOH 


1 


( 


\ 


S 

i. 2000 

Q. 


- 


\ 


o 

o 
o 

!000 


- 




n 


. ' '■ '," ".'.-' - . 1 .. 



9.5 



1 0.0 10.5 1 1.0 11.5 

pH 



12.0 



12.5 



Fig. 4.1 Color residual as a function of final pH 



8000 



.| 6000 

o 
o 



4000 



2000 - 







T 

9 




Color = 5000 

o= Mg^ = 300 mg./l - Ca{0H)2 


■ \ 


o^*»t— 


•=Mg* = 400 mg/l- NaOH 


1 


L 


f^^^"^ 7 1 1 



9.5 



10.0 10.5 1 1.0 

pH 



11.5 



12.0 12.5 



Fig. 4.2 Color residual as a function of final pH 



43 



initial color o£ 2500 or 5000. The constant magnesium dose used for 
each curve is indicated on the figures. The minimum pH at which 90% 
color removal was achieved was 10.6. The final color was dependent 
of the pH control agent. At pH 10 when CaCOH)^ was used, the final 
color of the waste was increased approximately 30% more than the ini- 
tial color. This did not occur with NaOH. The same degree of color 
removal was obtained from pH 10.6 to pH 11.4 using CaCOH)^ or NaOH. 
The degree of color removal decreases past 11.4 when NaOH was used. 
This did not occur with Ca(OH) . 

These figures show that a 90% reduction of the original color 
of the waste was first reached at pH 10.6 for each color. This indi- 
cated that the coagulation pH was independent of the variability in 
the waste color. From the data presented in Figures 4.1 and 4.2 it 
was concluded the coagulation pH was 10.6. 

4-1.2 Coagulant Dose 

A NSSC waste with a color of 2500 or 5000 was coagulated at pH 
10.6 with a varying magnesium dose. These data are presented in Fig- 
ures 4.3 and 4.4. The pH control agents were CaCOH)^ and NaOH. Mag- 
nesium coagulation of the NSSC waste with either pH control agent 
was able to remove 90% of the color. The required magnesium dose 
for the waste with a color of 2500 was 100 mg/1 when CaC0H)2 was used 
for pH control. The magnesium dose was 200 mg/1 when NaOH was used 
for pH control. The required magnesium dose for the waste with a 
color of 5000 was 200 mg/1 when CaCOH)^ was used for pH control, and 
was 400 mg/1 when NaOH was used for pH control. The required coagu- 
lant dose was directly proportional to the color of the NSSC waste. 



44 



The same color reduction was achi^eved when Ca(0H3 or NaOH was 
used for pH control. As is indicated in either Figure 4.3 or 4.4, a 
larger coagulant dose was required to achieve 90% color reduction when 
NaOH was used to control pH. The coagulant dose was approximately 
50% less whenCa(0H)2 wa^s used to control pH. 

4-1.3 Variation of Coagulation pH with Coagulant Dose 

All NSSC waste solutions were made by diluting a stored NSSC 
liquor with tap water. A waste color of 5000»was tested to deter- 
mine if the coagulation pH was dependent on coagulant dose. Figure 
4.5 shows for a color of 5000, the optimum color removal again occur- 
red at pH 10.6. The magnesium dose was 400 m^/1 and Ca(0H)2 was used 
to control pH. This was the same pH at which maximum color removal 
occurred at a color of 5000 for a magnesium dose of 300 mg/1 in con- 
junction with Ca(OH) . It was therefore concluded that the coagula- 
tion pH did not vary with coagulant dose. The only significant differ- 
ence between the curves in Figures 4.3 and 4.4 was the presence of the 
calcium ion when Ca(OH) was used for pH control. The stored NSSC 
liquor was made in a digestion process which used NaOH and contained a 
large concentration of sodium. The additional increase in sodium con- 
centration was therefore not significant when NaOH was used to adjust 
pH. The sodium concentration of wastes with colors of 2500 and 5000 
was 40,000 mg/1 and 80,000 mg/1 respectively. The amount of sodium 
increase when NaOH was added to adjust pH was always less than 1000 
mg/1, or less than 2.5%. Monovalent ions, such as. sodium, generally 
do not complex organic compounds to the same extent as divalent ions. 



45 



1000 



(O 



c 

3 



750 - 



o 
o 

k 500 



o 
o 



250 h 




Ca(OH) 



Color = 2500 
pH = 10.6 



100 



200 



300 
Mg mg/l 



400 



500 



600 



Fig. 4.3 Comparing NaOH and Ca(0H)2 for color 
removal via magnesium coagulation 



200C - 



Color = 5000 
pH = l0.6 



«I500 

'E 

3 

o 

o 

I 

s: 1000 

V. 

O 
O 
O 

500 



Ca(0H) 








lUO 



200 



300 
Mg mg/l 



400 



500 



600 



Fig. 4.4 Comparing NaOH and Ca(0H)2 for color 
removal via magnesium coagulation 



3000 



o 
V2OOO 



o 
o 
o 



1000 



J2 O. 



46 



4000 



Color- 5000 
Mg*^= 400 ma^l 
Ca(0H)2 



9.5 



10.0 10.5 I 




PH 



11.5 



12.0 



12.5 



Fig. 4.5 Verification of coagulation pH 



47 



The additional sodium added to the waste to adjust pH was not a 
significant increase in sodiim concentration, and did not extensively 
form any complexes. ' 

The total calcium concentration in the untreated waste was 
approximately 40 rag/1 as Ca"^*. When CaCOH)^ was used to adjust pH 
the calcium concentration in the waste was increased to 600 mg/1. 
This was significant because the calcium concentration increased 
and probably did extensively complex the organic compounds. 

Figures 4.3 and 4.4 show that the magnesium required to remove 
90% of the color was reduced when CaCOH)^ was used for pH control 
rather than NaOH. Both calciijm and magnesium are divalent ions and 
will form common complexes with the organic compounds in pulp wastes. 
When Ca(OH) was added to control pH, Ca*"* complexed many organics 
that Mg"^"^ would have normally complexed in the absence of the added 
Ca"^"^. Therefore Ca(0H)2 reduced the required coagulant dose. 

A magnesium dose of 100 mg/1 removed 90% of the color from a 
NSSC waste with a color of 2500 when Ca(0H)2 was used for pH control. 
A magnesium dose of 200 mg/1 was necessary for 90% color removal 
when NaOH was used to control pH. The complexing ability of the 
calcium ion was responsible for a 50% reduction in the coagulant 
dose. Since NaOH is more expensive than Ca(0H)2 and does not reduce 
the coagulant dose, CaC0H)2 was chosen as the pH control agent. 

4-1.4 Magnesium Remaining in Solution as a Function of Final pH 

The magnesium remaining in the treated waste after color removal 
as a function of final pH is presented in Figure 4.6. For both 



48 



E 





' 


- 




< 
300 


\ 




Ca(0H)2 

• Color = 2500 


2 50 


A 




Color = 5000 


200 


- 






( 
150 


- 


l^ \ 




100 


- 






50 




1 '. 


1 LJ l.„ 



10.0 



las 



II.O 



11.5 



12.0 



pHf 



Fig. 4.6 Magnesium remaining in solution as a 
function of final pH 



49 



colors tested, 35-40 mg/1 of magnesium remained in solution when the 
final pH was 10.6, The amount of magnesiirai remaining in solution 
was reduced to 4-10 mg/1 when the final pH was increased to 11.0. 
Magnesium in solution remained approximately constant past 11. Coagu- 
lation at pH 11 saved approximately 30 mg/1 of magnesium from being 
wasted in the treated effluent. The Ca(0H)2 dose was increased 
approximately 50 mg/1 to raise the coagulation pH to 11. The magne- 
sium saved was worth more than the Ca(OH) used to increase the pH. 
Therefore, it was decided to increase the coagulation pH to 11. 

Increasing the coagulation pH to 11 accomplished two things. 
First, enough magnesium was recovered to make the coagulation pro- 
cess less expensive. Second, the allowable fluctuation in coagula- 
tion pH was increased. The per cent color removed was significantly 
less at any pH less than 10.6. But when the coagulation pH was 11, 
a reduction of 0.4 pH units would not significantly affect color removal, 

4-1.5 Magnesium and Ca(0H)2 Dose as a Function of Initial Waste Color 

The waste effluent from a semichemical neutral sulfite plant 
varies in color intensity. Because of this variability, the amount 
of CaCOH) and magnesium to remove 90% of the initial NSSC color was 
determined as a function of the initial color of the waste. The 
Ca(OH)- required is presented in Figure 4.7. The magnesium require- 
ment is presented in Figure 4.8. Both the CaCOH) 2 and the magnesium 
requirements, were directly dependent on the initial color of the 
waste. This suggested a stoichiometric relationship between color 
and coagulant dose. 



50 



o 

o 

I 



o 
o 

o 









lopoo 


10 mg/l maximum Mg residual 
Lime mg/l = 750 t 0.10 (color) 


/ 


7500 


- 


/ 


5000 


/ 




2500 


sf 


1 I 







7500 



3 


5000 


O 

u 
1 

♦- 

a. 




o 
o 


2500 



500 1000 1500 2000 

Ca(0H)2 mg/l 

Rg. 4.7 Lime dose as a function of initial waste 

color for magnesium coagulation 



90% minimum color removal 
Mg mg/l = 0.060 (color) 



100 



200 



Mg^ mg/l 



300 



400 



Fig. 4.8 Magnesium dose as a function of 
initial waste color using lime 



51 



4-2 Waste Characteristics 

4-2.1 Untreated Waste Titration Curves 

The acid strength o£ the untreated NSSC waste was determined 
by titrating 50 ml samples with varying colors with 1.0 N H2SO4. The 
acid strength o£ the waste was defined as the milliequivalents of 
acid required to change the waste pH from 12 to 2. The results of 
these titrations are presented in Figures 4.9 through 4.13 for NSSC 
waste colors of 2,500 to 40,000. The NSSC waste was obtained from 
the Sunoco Products Corporation in Hartsville, South Carolina and is 
denoted as sodium base Sunoco NSSC waste in Figures 4.9 through 4.13. 
As the color of the NSSC waste was increased, the acid strength of the 
NSSC waste also increased. This indicated that color was acidic, and 
an increase in color wo'uld increase the acidity of the waste. 

The titration curves indicated the acidity of the waste was 
gained from two functional groups or mixtures of functional groups. 
These functional groups had pK values in the range of carboxylic 
acids and phenols or enols. The equilibrium constants for the NSSC 
wastes were approximated graphically and are presented in Table 4-1. 
The pK values were identified by locating the inflection points on 
the titration curves. Approximately 66% of the data points are not 
represented in Figures 4.9 through 4.13 in order that the titration 
curves would be uncluttered and clear. These points occurred in two 
areas, both of which were identified by slight humps on the titration 
curves. These pK values are approximately 4.6 and 9.8. They differed 
by four orders of magnitude, which was a large enough separation to 
allow graphical approximation of pK values. 



52 



14 - 



12 r 



PH 



8 - 



6 - 



4 - 







. Waste 


= Sodium 


base 


sunoco 


NSSC 


Vo = 


50 ml 








. Color 
» ^ 


= 2500 


^ 


/ 


_J 1 1 1 1 u 



10 



8 6 4 

meq acid X 10 



2 4 6 8 

meq base X 10 



10 



Fig. 4.9 Titration curve of sodium base sunoco NSSC waste 
with color equal to 2500 



53 



14 - 



12 - 



Waste = Sodium base 



PH 



10 
8 



6 - 



4 - 




8 6 4 

meq acid X 10 



2 4 6 8 

meq base X 10 



10 



Fig. 4.10 Titration curve of sodium base sunoco NSSC waste 
with color equal to 5000 



54 



14 



L Waste = Sodium base sunoco NSSC 



PH 



12 ' 



10 - 



8 - 



6 • 



4 ■ 



2 - 



10 




8 



6 4 

meq acid 



4 6 

meq base 



8 



10 



Fig. 4.11 Titration curve of sodium base sunoco NSSC waste 

with color equal to 10,000 



55 



14 
12 
10 



8 ■ 



Waste = Sodium base sunoco NSSC 
Vq = 50 ml 
Color = 20,000 



PH 




8 



6 4 

meq acid 



4 6 

meq base 



8 



Fig. 4.12 Titration curve of sodium base sunoco NSSC waste 

with color equal to 20,000 



56 



Waste = Sodium base sunoco NSSC 



PH 



12 - 



10 
8 



6 - 



2 





Fig. 4.13 Titration curve of sodium base sunoco NSSC waste 
with color equal to 40,000 



■TABLE 4-1 

GRAPHIC DETERMINATION OF pK^ OF 
SODIUM BASE NSSC WASTE 



57 




40,000 

20,000 

10,000 

5,000 

2,500 



9.6 

9.7 

10.0 

10.1 

9.5 



4.6 

4.5 
4.7 
4.7 
4.5 



Average 



9.7 



4.6 



58 



4-2.2 Comparison of Untreated and Treated Waste Titration Curves 

The acid strength o£ the untreated NSSC waste was compared with 
the acid strength of the treated NSSC waste. This was done in order 
to determine if any reduction in acidity occurred during the color 
removal process. The previous titration curves of the untreated waste 
revealed significant acidity in the weak and very weak acid range. 
These acids would be ionized at pH 11 and available to participate 
in a chemical reaction. Magnesium as Mg is a Lewis acid and is 
capable of reacting chemically with the ionized anions from the waste 
acids. 

The titration curves are presented in Figure 4.14 for the same 
NSSC waste before and after treatment. The acid strength of the 
NSSC waste was reduced by the color removal process. Before treat- 
ment, 0.8 meq of base was required to titrate the waste from pH 12 
to pH 9. After treatment, only 0.4 meq of base was required to pro- 
duce the same change in pH. The amount of base to change the waste 
from pH 12 to pH 3 before and after treatment was 1.65 and 1.0 meq 
respectively. Very weak acids have pK values ranging from approxi- 
mately 8 to 10. Weak acids have pK values of approximately 3 to 5. 
The total reduction in acid strength during the color removal pro- 
cess was 0.65 meq. Approximately 0.4 meq of this reduction occurred 
in the very weak acid range from pH 12 to pH 9. This was approxi- 
mately 60% of the total reduction in acid strength. From the titra- 
tion data, it was concluded that color reduction by magnesium coagu- 
lation does result in at least a reduction of the acids present in 
the NSSC waste. 



59 



PH 



6 
5 
4 



Jreoted NSSC 

waste 
■Color = 485 
Vol=50ml 



Untreated NSSC waste 
Color = 5000 
Vol = 50 ml 




10 
meq X 10 



15 



Fig. 4.14 Titration curve of treated and 
untreated NSSC waste 



60 



4-2.3 Waste Content 

The treated and untreated waste analyses shown in Table 4-2 
were done by the United States Air Force Environmental Health Labora- 
tory at Kelly Air Force Base in Amarillo, Texas. The xintreated waste 
was a sodium base NSSC waste that was prepared from a stored liquor 
obtained from the Sunoco Products Company. Hydrochloric acid and sul- 
furic acid were used to adjust the pH to 7.6 before shipment. 

4-3 Color Removal Mechajiism 

4-3.1 Color and Magnesiiim Titration Curves 

A volume of 200 mis of NSSC waste was dosed with 80 mg (400 mg/1) 
of magnesixm as Mg . This solution was titrated with a 1.0 N NaOH 
solution to a pH of 12.0. Tne titration curve for this experiment is 
presented in Figure 4.15. In the first portion of the curve, an 
inflection point was present at pH 9.6. This was approximately the 
second pK determined earlier from the NSSC waste titration curves. 
The solution was slightly buffered by the colored NSSC waste at this 
point in the titration. At pH values higher than 9.6, the acids in 
the NSSC waste were ionized. 

In the second portion of the curve, another inflection point 
was found at pH 10.8. Coagulation and 90% color removal occurred at 
all pH values greater than or equal to 10.6. At pH 10.8 in the titra- 
tion curves, magnesium was acting as a buffer by hydrolyzing and pre- 
cipitating out of solution. Color removal was accomplished when the 
buffering capacity of the colored waste and the magnesium were excee- 
ded. The acids were ionized and were capable of an acid-base reaction 



61 



. TABLE 4-2 

UNTREATED AND TREATED NSSC WASTE ANALYSIS. TREATMENT WAS 
WITH 150 mg/1 Mg"""*" AND Ca(0H)2 TO ADJUST pH TO 11.0. 



Item 



Lab Analysis 
(mg/l unless noted) 



Untreated 



Treated 



1. 


Color - Pt-Co units 


2000.000 


150.000 


2. 


Turbidity - JTU's 


3.000 


4.000 


3. 


Chemical oxygen demand 


1510.000 


784.000 


4. 


Total suspended matter 


0.000- 


0.000 


5. 


Volatile and fixed 








suspended matter 


0.000 


0.000 


6. 


Oils and greases 


0.800 


0.500 


7. 


Surfactants - as mg/1 LAS 


0.800 


1.600 


8. 


Chlorides 


56.000 


920.000 


9. 


Flourides 


1.100 


0.500 


10. 


Phosphates 


0.500 


0.300 


11. 


Sulfates 


200.000 


1150.000 


12. 


Cadmium 


.01 


.02 


13. 


Chromium (hexavalent) 


.01 


.01 


14. 


Chromium (total) 


,03 


.05 


15. 


Copper 


.05 


.03 


16. 


Cyanides 


.01 


.02 


17. 


Iron 


.72 


.1 


18. 


Lead 


.05 


.07 


19. 


Manganese 


.28 


.05 


20. 


Silver 


.01 


.02 


21. 


Zinc 


.1 


.05 


22. 


Mercury 


.005 


.005 


23. 


Total organic carbon 


. 530.000 


350.000 


24. 


Nitrite nitrogen 


.06 


.02 


25. 


Ammonia nitrogen 


.8 


.2 



62 



PH 



Buffering due to magnesium 




Initial color = 5000 
Vol = 200 ml 

Mg** = 400 mg/l 
I N NaOH 



Slight buffering due to color 



1 1 I 



2 4 6 8 10 12 14 16 

meq 

Fig. 4.15 Titration curve of raw waste dosed with magnesium 



63 



with magnesium. This reaction could involve the color in a formation 
of an insoluble precipitate. The removal of this precipitate would 
remove the color bodies. 

4-3.2 Magnesium, Calcium, Color and Organic Carbon Residuals After 
Coagulation 

A NSSC waste with an initial • color of 2,500 was coagulated with 
a constant magnesium dose of 150 mg/1. The final pH of coagulation 
was varied from 10 to 11.5 using Ca(0II)2 to control pH. The magne- 
sium, calcium, organic carbon and color residuals were determined 
after coagulation. The total organic carbon concentrations and color 
intensities after coagulation are presented in Figure 4.16. The mag- 
nesium and calcium concentrations remaining after coagulation are 
presented in Figure 4.17. 

The total organic carbon concentration was reduced 40% when the 
NSSC waste was coagulated at pH 11.5. The residual color at this 
point was 157 Pt-Co color units. When the waste was coagulated at 
any pH from 10.6 to 11.2, approximately 34% of the total organic car- 
bon was removed. The average residual color in this pH range was 
197 Pt-Co color units. Increasing the coagulation pH to 11.5 would 
only remove an additional 1.8% of the initial color. Coagulation at 
any pH from 10.6 to 11.2 removed 92% of the initial color. 

When the waste was coagulated at pH 11.5, an additional 6% 
(27 mg/1) of organic carbon was removed. As was noted, the additional 
color reduction at pH 11.5 was 1.8%. When the waste was coagulated 
at any pH from 10.6 to 11.2, the organic carbon was reduced 154 mg/1 



64 



400 ■ 



_300 

>^ 

o> 

E 

q 

S 200 



100 



9.5 



Initial Concentrations 
T.0.C.=455 mg/l 
Color= 2500 




100 



10.5 



11.0 



11.5 



pHf 



Fig. 4.16 Organic carbon and color residuals 
as a. function of finol pH 



3000 



c 

3 



2000 5 



o 
o 
O 



- 1000 



12.0 



150 - 



100 - 



E 



50 - 



■ •\ 


\ 




Initial Concentrations 

Mg* = !l mg/l 
^ 00^^ = 15 mg/l 


- 


\ 


y 


y^Z^ remaining 


^ 


A 


\f 




\ 


(Vig remaining 






u 


^*^^....^^_^ • 


1 1 



'9.5 



10.0 



Fig. 4. 1 7 



10.5 



pH 



11.0 



f 



11.5 



Magnesium and calcium residuals 
as a function of final pH 



Coagulant dose = 150 mg/i Mg 



Vt 



■ 600 



400 



- 200 



E 
o 



12.0 



65 



and 2,300 Pt-Co color units were removed. In this pH range, color 
was reduced by 15 Pt-Co color units for every mg/1 of organic car- 
bon removed. For each additional mg/1 of organic carbon removed by 
coagulation at pH 11.5, the color was reduced by only 1.5 Pt-Co 
color units. From these data it was concluded that not all the 
organic carbon in the waste contributed equally to the waste. color. 
The residual color increased 32% when the coagulation experi- 
ment was attempted at pH 10.0. The magnesium residual curve in 
Figure 4.17 shows that no magnesium precipitated out of solution at 
pH 10, All of the magnesium was therefore available to form che- 
lates with the NSSC waste. Calciirai ions causing increases in the 
color of a kraft waste due to chelation were reported by Luner and 
Dence (1971), Color increasing chelates formed by magnesiiom and 
quinones have been reported by Day and Underwood (1967) , Aromatic 
quinones are an integral part of basic lignin structure, and lignin 
is responsible for color in pulp waste. The color increase at pH 
10 was probably due to the chelation of lignin building units, possi- 
bly direct chelation with quinones. 

None of the 150 mg/1 of magnesium was removed after coagulation 
at a final pH of 10.0. The total amount of magnesixom available for 
coagulation was the magnesium dosage and the magnesium present in the 
waste. For the data presented in Figure 4.17, the total amount was 
161 mg/1 magnesium. When magnesium precipitation began, a correspon- 
ding drop in color intensity was observed, as was a corresponding 
drop in organic carbon. The concentrations of magnesium, organic 



66 



carbon, and color decreased with increasing pH, indicating that the 
decreases in these three parameters were related. Corresponding 
decreases in magnesium, color and organic carbon occurred simulta- 
neously,. A possible relationship for these simultaneous reductions 
could be a chemical reaction between the color producing organic com- 
pounds and the magnesium ions. This relationship would result in the 
chelation and precipitation of a magnesium organic color-body complex. 
A second possibility could be the adsorption of the chelated organics 
onto the voluminous magnesium hydroxide floe." The data presented in 
Figures 4.16 and 4.17 could conform to either of these postulated 
mechanisms . 

Ca(OH)_ was used as the pH control agent in these experiments. 
As the Ca(OH)„ dissolved, the pH and calcium concentration increased. 
The increasing pH probably caused the solution to become supersatura- 
ted with respect to the magnesium-color body compound. This hypothe- 
sis is supported by the data presented in Figure 4.17. As the pH 
increased from 10.0 to 11.5, the calcium in solution increased from 
237 to 681 mg/1 as Ca . The magnesium in solution decreased from 
161 to 7 mg/1 as Mg . This increase in calciiom and decrease in mag- 
nesium concentrations can be visualized as a reaction between the 
dissolved Ca(0H)2 and Mg"'"''. Such a reaction is shown in Equation 4-1. 
The value of AG° is -12.601 kcal/mole with the reaction proceeding 
from left to right. 

Ca"^"^ + 20H" + Mg"^"^ = Mg(0H)2(s) + Ca** (4-1) 
Total organic carbon, magnesium and color residuals were 



67 



determined after coagulation with a varying magnesium dose. In these 
experiments, the pH was held constant at 10.6. The pH controls used 
were NaOH and Ca(OH) . The data from these experiments are presented 
in Figures 4.18 and 4.19. In these figures the residual color, organic 
carbon, and soluble magnesium are plotted as functions of the total 
millimoles of magnesium available. The total millimoles of magnesium 
available consists of the initial magnesiim plus the coagulant dose. 

The initial color of the NSSC waste used to plot Figure 4.18 is 
approximately half that of the initial color of the waste used to plot 
Figure 4.19. The scales in Figure 4.18 are one-half the scales in 
Figure 4.19. This was done so the two figures could be directly com- 
pared without being misleading. 

In both figures, a decrease in organic carbon was accompanied by 
a decrease in color removal. The beginning of floe formation is iden- 
tified by the dashed lines in Figures 4.18 and 4.19. This point was 
identified when the floe became visible to the naked eye. No magne- 
sium was removed from solution until floe formed. Once floe formed, 
the color was reduced below the initial color level of the waste. 
Before this point, a color increase had occurred due to the chelation 
of magnesium and calcium with the NSSC waste. After formation of 
floe, the color, magnesium and organic carbon concentrations were 
decreased. The floe formed at a smaller magnesium dose using CaC0H)2 
compared to using NaOH for two reasons. First, the initial waste 
color treated with NaOH was higher. Floe was formed with Ca(0H)2 when 
the total amount of available magnesium was 2.19 millimoles. This 
point was not reached with NaOH until the total millimoles of Mg 



6S 



3000 



2500 



I 2000 



o 

o 
I 



o 
o 

o 



1500 



1000 



500 - 



■ 500 


- 


Vol= 1 liter 
pH= 10.6 


- 




o = Color remaining 

■H- 

D = Mg remaining 


375 


\ •^'^--. 


• • = T.OC. 


o 






~ "^^ 


A • 


' ^^-.....^^^^ 


en 
E 


•i\ 


-^^..^^^^ 




d \ 


^'""••s^ 


d 


,^\ ■ 




- g 250 




^^ 


" 


ill \ 


^y^^ 


■ 125 


' ye |^^~^^--a 


-^---^"^^ 






>^..___^ 




1 1. ... . .... .. 


1 1 1^_ 



- 75 



E 



50 : 



c 
c 

E 
a> 



25 



2.5 5.0 7.5 

Total m mo! Mg available 

Fig. 4.18 Color, T.O.C, and Mg"^ residual after Mg^ 
coagulation using Ca(0H)2 for pH control 



10.0 



69 



6000 - 1000 - 



5000 



^4000 



2000 



1000 



750 



c 

3 




o 




O 








o 
1 




cn 




a: 


3000 


- b 


500 


k. 




d 




o 

o 




d 




o 




h- 





250 - 



Vol = I liter 

pH = 10.6 

o = Color remaining 

D = Mg^' remaining 

• = lO.C. 




150 



E 



100 : 



c» 



c 
"c 
'o 
£ 

OC 



- 50 



10 



•tt 



20 



Fig. 4.19 



Total m mol Mg available 
Color, T.O.C., and Mg residuals after Mg 



coagulation using NaOH for pH control 



70 



available were 9.46. If the initial waste color was the only reason 
for the larger magnesium requirement with NaOH, then the amount of 
available magnesium required to form floe would increase in propor- 
tion to the color increase. This was not the case. The initial 
color of the waste treated with NaOH was approximately twice that 
treated with Ca(0H)2- If the magnesium dose was directly propor- 
tional to the initial color, then extrapolating the Ca(0H)2 treat- 
ment dose would give 4.38 mM/1 as the necessary magnesium dose for 
the NaOH treatment. It was found that 5.08 mM/1 more of magnesium 
was necessary to form floe using NaOH. This difference was due to 
the presence of calcium ions when Ca(0H)2 was used to control pH. 
Floe formation did not occur with the first magnesium additions with 
either pH control agent. Calcium ions from Ca(0H)2 complexed some 
of the organics in the waste that would have been complexed by the 
magnesium had Ca(0H)2 not been used. This enabled floe formation 
to occur at a smaller magnesium concentration. 

Two different cliemical reactions, chelation and precipitation, 
have been identified in the color removal process. First, chelation 
occurred between the divalent metal ions and the ligands present in 
the NSSC waste. The chelation demand of the NSSC waste was satisfied 
before color removal occurred. The magnesium chelates were partly 
reduced by using Ca(0H)2 to adjust pH. After the chelation demand 
was satisfied, floe formation and color removal occurred as shown in 
Figures 4.18 and 4.19. 



71 



4-3.3 Stoichiometry of Color Removal from NSSC Waste by Mangesiv un 
Coagulation 

A possible mechanism of color removal was adsorption of the color 
bodies on magnesium hydroxide floe. The colored organics would not 
have been involved in a chemical reaction that formed a magnesium com- 
pound, but would have become attached to the floe by Van der Waals 
forces or hydrogen bonding. If magnesium hydroxide floe were formed, 
two moles of hydroxide would be required for every mole of magnesium 
removed from solution. If an insoluble precipitate formed that was 
a chemical compound consisting of magnesium, feydroxide and organic 
ions, the moles of magnesium ions removed divided into the moles of 
hydroxide ions removed would be less than two. 

The moles of magnesium removed divided into the moles of hydroxide 
removed is presented as a ratio in Figure 4.20. There are three dif- 
ferent experiments represented in Figure 4.20. In the first two, 
CaCOH)^ was used to control pH for color removal from NSSC wastes with 
initial colors of 2500 and 5000. In the third experiment, NaOH was 
used to control pH for color removal from an NSSC waste with an ini- 
tial color of 2500. 

The moles of magnesium removed were found by measuring the mag- 
nesium concentrations before and after color removal. The moles of 
hydroxide removed were found by difference. First the moles of 
hydroxide necessary to raise the pH to 10.6 were found. Then this 
amount was subtracted from the moles of hydroxides required to raise 
the pH to 10.6 after the magnesium dose was added. The difference 
was the hydroxide . demand of the magnesium used to Qoagulate the 
color, and was represented as [ohJj^ ++ in Figure 4.20. If. NaOH was 



72 



-H- 



8 



3 
o 






2 
I 





500 coagulated by Mg using lime 

o = Initial color 5000 coagulated by Mg* using lime 

■ti- 



© = initial color 5000 coagulated by Mg using NaOH 




10 



15 



20 



25 



fr 



mmol Mg available 

Fig. 4.20 Ratios of [Oillivig-"-/ Qvlg^l 

[OH]w it is the moles of hydroxides 
required by the magnesium for color re- 

Mg J is the moles of Mg 
required for color removal. 



73 



used to control pH, the moles o£ hydroxide were measured from the 
direct addition of a 1.0 N NaOH solution. If Ca(0H)2 was used to 
control pH, the increase in calcium concentration before and after 
coagulation was measured. The calcium increase was doubled to deter- 
mine the moles of hydroxide required in the color removal process. 

The curves shown in Figure 4.20 represent magnesium to hydroxide 
molar ratios for the floe formed in the color removal process. The 
coagulation pH was 10.6. Both Ca(0H)2 and NaOH were used to control' 
pH. All of the data points in Figure 4.20 represent some degree of 
color removal . 

When NaOH instead of Ca(0H)2 was used to control pH, a greater 
magnesium concentration was required before any floe was formed. 
This was due to chelation and is shown by the separation of the two 
curves for wastes of equal initial color in Figure 4.20. 

The initial points on each of the curves in Figure 4.20 repre- 
sent the beginning of floe formation. As color removal and floe for- 
mation increased, the curves eventually stabilized at 1.5. At the 
low magnesium doses used initially this ratio was not stable because 
of the chelation demand of the waste. Once this demand was exceeded, 
the ratio stabilized at 1.5 and remained there for all subsequent 
magnesium doses. 

A ratio of 1.5 hydroxide ions to 1.0 magnesitmi ion does not pro- 
duce an electrically neutralized compound. Another anion had to con- 
tribute one-quarter of the total negative charge for the precipitate 
to be electrically neutral. Color bodies are negatively charged and 
color was removed as magnesium ions were precipitated. If the color 



74 



bodies were involved in a chemical reaction with the magnesium and 
hydroxide ions, then the molar ratio would be less than two. The 
negatively charged color bodies would electrically neutralize the 
precipitate. The ratio of 1.5 indicates that an insoluble precipitate 
was formed in the ratio of 30H~ : 2Mg : 1R~ where R represents the 
color body. 

The formation of an insoluble precipitate is further supported 
by the stability of the ratio. If the magnesium-color body complex 
became enmeshed in a Mg(0H)2 precipitate, the overall OH/Mg ratio 
would be less than two. Some of the magnesium removed would be attri- 
butable to the enmeshed chelate and some to the Mg(0H)2 floe. The 
ratio, however would not be stable for an increasing coagulant dose. 
As the coagulant dose would increase, the ratio would approach two 
because mostly Mg(0H)2 would be formed after the chelation demand of 
the waste was satisfied. As shown in Figure 4.20, the molar ratio 
does not vary after becoming stabilized at 1.5. From the data pre- 
sented, it was concluded that a chemical reaction was the mechanism 
by which color was removed from solution. 

Calcium hydroxide has been used successfully to remove color from 
a kraft waste at a pulp plant in Riceboro, Georgia. Dissolved Ca(0H}2, 
used in magnesium coagulation, could precipitate as Ca(OH)-, or some 
other compound and remove NSSC color bodies. However, based on solu- 
bility product calculations, no Ca(0H)2 would precipitate at the 
Ca(OH)„ doses and pH required by magnesium coagulation. 

An experiment was done to determine the color removal capability 
of Ca(0H)2 with reference to a NSSC waste. The residual color and pH 



75 



were determined in a NSSC waste after Ca(0H)2 addition. These results 
are shown in Figure 4,21. An increase in the residual color of the 
waste was noted for the initial doses of CaC0H)2- This was due to 
chelation of calcium ions with the organic compounds in the waste. 
The maximum Ca(0H)2 dose was 2000 mg/1, with a resultant pH of 12 and 
a color reduction of approximately 45%. No. color was removed until 
the pH was 11.2, The pH used in magnesium coagulation was 11. At 
pH 11 the use of Ca(0H)2 alone slightly increased the residual color 
of the waste. This is shown in Figure 4.21. These results indicate 
that CaC0H)2 does not remove any color in the magnesium coagulation 
process when the coagulation pH is 11. 

4-4 Settling of Coagulated Wastes 

4-4.1 Purpose of Settling Tests 

The purpose of the settling tests was to minimize the voliime of 
sludge and gain some knowledge of the factors governing the settling 
process. The Sludge Volume Index, SVI, was determined after each 
settling test on all of the sludges produced during coagulation. 

The settling tests were conducted as described in Section 3-2.8 
of Chapter 3. Settling tests were performed on the coagulated wastes 
and on polymer treated coagulated wastes. Cationic, nonionic, and 
anionic polymers were used as settling aids in the tests. All of 
the polymers tested were supplied by the American Cyanamid Company. 

Since the magnesium dose depended on the initial color of the 
waste, a constant concentration of suspended solids was produced in 



76 



500 



Fig. 4.21 



1000 1500 

Ca(0H)2 mg/l 



2000 



Color and fxH of a NSSC waste as 
a function of Ca(0H)2 concentration 



- 5000 



- 4000 




77 



all wastes with the same initial color. The use of polymers as a 
settling aid was found to have a negligible effect on the suspended 
solids produced. A coagulated waste with an initial color of 5000 
was found to have a suspended solids concentration of approximately 
1800 rag/1. If the initial color was reduced by half, the suspended 
solids produced by coagulation also were reduced by half. 

4-4.2 Sludge Settleability 

The type, functional group, charge and approximate molecular 
weight of the polymers used in the settling tests is presented in 
Table 4-3. The Sludge Volume Index is also presented in Table 4-3 
as a function of polymer dose and polymer type. 

The SVI of the raw sludge was 352. This was increased to approxi- 
mately 550 when a cationic polymer was used as a settling aid. The 
floe was still completely suspended after 30 minutes. No sludge- 
supernatant interface had developed when any concentration of cationic 
polymer was added. The floe was very small, completely dispersed and 
appeared to be in a state of compression during the entire settling 
test. 

When a nonionic polyacrylamide was added to the sludge, the SVI 
was reduced. When 5.0 mg/1 of a nonionic polymer was used as a sett- 
ling aid, the SVI was reduced to 178. The physical appearance of 
the floe changed very little. It appeared very small but the degree 
of dispersion was less than the dispersion of the raw sludge. The 
floe appeared more dense. The high molecular weight and large size 
of the nonionic polymer was effective in consolidating the floe. 



78 



TABLE 4-3 



POLYMER DESCRIPTION AND SVI FOR POLYMER ASSISTED 
SLUDGES PRODUCED FROM AN INITIAL COLOR OF 5000, 



++ 



Mg = 350 rag/l, Ca(0H)2 = 1500 mg/1, pH = 10.8 



Polymer Type 




Subunit 






Charge Molecular 














Weight 


575C Cationic 


Amine 








High 




500,000 


1905N Nonionic 


Polyacrylamide 






Zero 


15 


,000,000 


1838A Anionic 


Polyacrylic Acid 






High 


15 


,000,000 


837A Anionic 


Hydrolyzed Polyacrylamide 


5% 


Low- 


15 


,000,000 


835A Anionic 


Hydro 


lyzed Polyacrylamide 


25% 


High 


15 


,000,000 


Polymer 




Sludge Volume Ind 


ex 






Dose mg/1 


















None • 


575C 


1838A 


1905N 


837A 


835A 


00 


352 














.03 






544 


342 




408 




.05 






547 








364 


.10 




544 












.30 






• 


362 






303 


.50 




547 








250 




1.00 
















1.50 








203 






198 


1.80 






422 










3.00 




547 








97 




5.00 




. 


294 


178 




81 


83 


10.00 




547 


294 






67 




15.00 






275 


% 




75 




20.00 






233 






67 




25.00 






230 






64 





79 



The smaller size molecular weight of the cationic polymer was ineffec- 
tive in consolidating the floe. The smaller polymer created repulsive 
forces among the floe particles, probably due to its size and positive 
charge. The nonionic polymer was larger and not charged. Increased 
floe settleability resulted from the polymer-floe interaction. 

Three anionic polymers were investigated as settling aids. The 
first of these was a colloidal polyacrylic acid (1838A) which required 
an activator before use. Once activated, the polymer formed small 
spheres approximately 0.5 mm in diameter. The polyacrylic acid 
decreased the SVI to 230 at a concentration of 25 mg/1. This was 
a significant reduction in SVI but an excessive polymer dose was 
required. The available surface area for floe interaction was much 
less when the activated spheres of polyacylamide were formed. The 
nonionic polymer was dosed as a clear liquor. It was completely 
soluble in the coagulated mixture and rendered more available sur- 
face area to the floe. 

Two additional negatively charged polyacrylamide polymers were 
investigated as settling aids. These polymers were added to the 
coagulated mixture as clear liquors and were completely soluble. 
A large amount of polymer surface area was available for floe inter- 
action. Within one minute of the 30 minute settling test for a 
polymer dose of 5.0 mg/1, the sludge volume had been reduced 85% with 
either of the polyacrylamides. The floe changed from a light well- 
dispersed floe to a heavy dense floe. The average size of the floe 
particles changed from approximately one micron to approximately 
one centimeter. 



80 



The interaction between the anionic polyacrylamide and the floe 
was quite rapid. The rapid interaction is shown in Figure 4.22. The 
sludge interface is presented as a function of settling time. The 
solids concentration in the sludge was increased approximately seven- 
fold due to the addition of 3.0 mg/1 of a 5% hydrolyzed polyacryla- 
mide. 

The negatively charged polyacrylamide was the most effective 
settling aid. A high degree of negative charge was not required on 
the polymer. This was shown by the identical effectiveness of 837A 
and 835A. A polymer is negatively charged by hydrolysis. The grea- 
ter the degree of hydrolysis on the polymer, the greater the polymer 
charge. The 837A polymer was 5% hydrolyzed, and the 835A polymer 
was 25% hydrolyzed. A 5% hydrolyzed polyacrylamide means that 95 out 
of every 100 monomer units are uncharged acrylamide groups; the 
remaining 5 monomer units are negatively Charged acrylic groups. 



Uncharged polyacrylamide: 



-CH2-CH- 



C-0 



NH- 



n 



Charged polyacrylamide: 



-CH^-CH-CH^-CH- 
2 I 2 , 

c=o c=o 

/ / 

0" NH^ 



81 



E 
I 


600 


x: 




f 


500 


o 
a 
«*- 

w 
C 


400 


0) 

■D 


300 



to 



200 



100 



Color = 5000 
Mg^ = 350 mg/i 
Lime = 1500 mg/i 




N Polymer 




-3.0 mg/i Hydrolyzed Polyacrylamide 
' Anionic Polymer 



20 30 

Settling time— minutes 



40 



50 



Fig. 4.22 Sludge settling velocity for polymer assisted and raw sludge 



t^<Sflai.^^=- i^^~ • 



82 



The completely uncharged polyacrylamide was not as effective as either 
charged polyacrylamide indicating the need for a negatively charged 
polymer during sedimentation. This need was met by a small degree of 
hydrolysis. Some degree of interaction between the charged carboxylic 
functional group and the magnesium floe was necessary for optimum 
settling. 

4-4.3 Mechanisms of Sedimentation 



There are two main areas of thought about the mechanisms of 
destabilization of colloids. One area deals with the colloidal sta- 
bility introduced through the mutually repulsive electrical double 
layers present on similar colloids. The electrical charge on the 
colloid surface will attract counterions, and if a sufficient number 
of counterions are available, colloidal destabilization or sedimen- 
tation will result. When this occurs the Van der Waals forces of 
attraction overcome the electrostatic repulsion and the colloids 
then can agglomerate and settle. However, there are many possibili- 
ties where the electrostatic energy involved in a colloid-counterion 
interaction will be far less than the energy from chemical bonds 
between colloid-coagulant interactions. Lamer et al. (1967) have 
developed a bridging theory in which polymers of high molecular 
weight can destabilize colloidal suspensions. If the polymer con- 
tains chemical groups which can interact with the colloids, then the 
polymer can destabilize the colloids. Once the colloids begin inter- 
acting with the polymer, a bridge is formed between the colloids by 
the polymer. As an increasing number of colloids become attached to 
the polymer bridge, the likelihood of destabilization increases. For 



83 



colloidal destabilization to occur by this model, the polymer dose 
must be coordinated with the colloidal concentration. It is possi- 
ble to restabilize a colloidal suspension by too great a polymer 
dose or by shearing the polymer with too high a mixing energy. There 
are many instances in wastewater treatment where negatively charged 
colloids are destabilized by anionic polymers. This phenomena 
can be explained by an interaction between the functional groups and 
the colloids, as in the bridging model. 

The electrophoretic mobility was measured on the floe particles 
to determine if the settleability of the floe particles increased as 
the floe charge decreased. The Helmholtz-Sraoluchowski (H-S) formula 
was used to determine the electrophoretic mobility. Riddick (1974) 
specified the applicability of different zeta potential formulas 
based on normality of suspending solution and particle diameter. He 
recommended the Helmholtz-Smoluchowski formula to measure the electro- 
phoretic mobility of any particle suspended in a 1 . ON solution whose 
diameter was 0.8 microns or greater. The floe particles produced in 
the NSSC waste by coagulation met these specifications. 

The H-S formula for determining zeta potential is as follows: 

ZP = 113,000(V^/D^)EM (4-2) 

ZP = Zeta potential in millivolts 

EM = Electrophoretic mobility in microns cm/sec volt 



V = Viscosity of suspending liquid at a given 
temperature in poises 

D^ = Dielectric constant of the suspending liquid 






84 



Figures 4.23 and 4.24 present the SVI and zeta potential as a 
function of polymer dose for a nonionic and anionic polymer. It 
will be shown later that a negative potential occurred on the floe 
as it was formed in the absence of any polymeric settling aid. If 
electrostatic reduction was the major mechanism of enhanced settling 
of the negatively charged floe, then the cationic polymer would have 
been the most effective settling aid. As Table 4-3 shows, the cat- 
ionic polymer stabilized the floe and severely hindered settling. 
Conversely, the anionic polymer was seen to be an effective settling 
aid. Restabilization of the floe was not achieved at the polymer 
doses tested. The zeta potential was observed to increase from -13 mv 
to -10 rav when the polymer dose was varied from to 5.0 mg/1 of 837A. 
It did not approach zero although the SVI of the sludge changed from 
352 to 83. The total change in ZP as settleability increased indi- 
cated that decreasing electrostatic repulsion was not the major 
mechanism for floe destabilization. The controlling mechanism was 
probably polymer bridging. 

For clarification of the coagulation reaction between magnesium 
and NSSC waste, eleetrophoretic mobilities were determined on magne- 
sium floe produced at varying pH's in tap water and in NSSC waste. 
The data for these experiments are presented in Table 4-4. The zeta 
potential as a function of pH is graphed in Figure 4.25. A magnesium 
concentration of 350 mg/1 was used to produce floe in both the tap 
water and the NSSC waste. 

The zeta potential of the magnesium hydroxide floe produced in 
the tap water was positive, and increased with increasing pH. The 



85 



^^ 

il <» I I aiH i 1 1 



o 
I 



o 


o 


o 


O 


O 


o 


o 


o 


O 


O 


in 


t 


ro 


CNJ 


_ 



o 



Q 

E 

O I 
rd <c 
w 
o 



O 



0) 

£ 
o 

CL 



I 

o 
a. 



10 

> 






C 

o 
> 

CO 



E 

a 



o 



c 
o 



o 

N 



o 

TJ 



£ 



(VJ 



< 

ro 
00 



SJIOAJIIJUJ -|D!;U8|0d d;3Z 



xspui euiniOA aSpnis 




o 
I 



o 

CM 

I 



O 

o 

in 




o 




1 


Jm 


m 




O 


F 






o. 


1 






en 




o 




> 


o 


^ 




o 


c 






V- 


o 




•^ ^. 


c 


f- 




D> 


H) 


o 




E 


o 


c 


o 
to 


1 


CL 

o 


O 




M 




o 




O 


N 


<4- 

0) 






T3 


CO 


o 


(U 


c 


q 


OJ 


E 


a 


a 












o 


^ 


0) 




CL 


fc 


o 




(VJ 





2 

m 
o 



S||OAj||jUi -|Dj^U9;0d D43Z 



xgpu] 9Uin|0A sBpnis 



86 





Cfl 




w 


Q 


Q 


m 


O 


u 


OS • 


3 


H PJ 


Q 


U H 


O 


PJ CO 


b: 


J < 


Dh 


PJ s 


pj 


z u 


a 


m c/3 


o 


PJ c/:) 


3 


S 2 


J 


f-H 


CO 


W 2 




CO t-H 


g 


B u 


1— 1 


uo 


en 


in 


M 


O <N 


Z 


1—1 


o 


^ 


<: 


oi 


s 


■ PJ 




I H 


Oi 


PL,< 


o 


S 


[I- 


u 


"!a- 


2 (X 


1 J 


hH < 


't < 


>■ H 


h- 1 


Di 


C4 H 


< 2 


J 2 


> M 


ca aj 




< H 


H u 


H O 


<:o 


a. 


un 




PJ • 


< 


H to 


H 


C/3 CN 


w 


< 


M 


S PJ 




o; 


Q 


CO H 


•^ 


CO <; 




2 cS 


>- 


PJ 


H 


Q O, 




2 S 
< PJ 


l-H 


H 


CQ 


a; 


O 


W PJ 




H U 


Q 


< < 


Di 


S Di 


E- 


PJ 


U 


&, > 


w 


<; <: 


ij 


H 


PJ 






2 





s 






o 


t^ 




c 


o u 


^ 


o 


>o 


PJ 


u 


LO 




o 


U CM 




■H 


<u 




s 


U5 



u 

a, >o 

tSl S Lrt 
CM 



<U S O 

> -H <D 



71 

(D 



Oh 

I 
C/3 



-o o 

O +-> -d 

.-I o C 

t-H 0) Hi 

O JH 

U -H 
Q 






13 

0) 

(D 1/5 t/l 

> 

C 



C 
O 

u 

•H 

s 



o 
> 



E 




P< 




0) 




o o 


t/l 


•H C 


o 


M-i nJ 


A 


•H 4-> 


s 


O U 




01 p 


o 


OhX) 


f-l 


c/5 C 


o 


o 


•H 


u 


£ 



• T3 
X -H . 
O O Cfl- 
f-l 1— I -H 

Oh O 

< u 



+-> 



CTi 

E- 



\D vO 1— t LO LO to "* 
to O CTi -* to to fO 



1— I t— I O O O O O 
+ + + + + + + 



■^ to 1— I t~^ r^ 1— I "* 
[~- 00 00 00 LO to ■* 



t^ ^0 i-H LO ^ 'd" 'd" 



to LO CTi \0 LO r— I CTl 
•^ LO 'y ^ 00 (N r— I 



un Lo r^ o) CT) oi CN 



o o o o o o o 
'y ■* to to to ^ -y 
+ + + + + + + 



t^ r-^ r^ o o o o 
\0 ^ ^ o o o o 



LO LO CN O CO \o ■«* 

CN I— I .— I t— I O O O 



o o o o o o o 

o o o o o o o 

o o o o o o o 

^ ^ -nJ- >* ^ ^ •* 



■^ to >— I O^ LO -* CN 



1— I r— I r-l O O O O 



in 



C/D 
C/5 

2 



rj- O^ \0 to t->. (N to 

O^OvOvO^OOOCTlO 



O O O O O O i-H 

I I I I I I I 



OLnOr-HCNLOO-^ 

to O vO CM to o •^ 



00 CTi 00 00 >— I (N to 

I I I I 1— I rH rH 

I I I 



'y -vf rH O Cvl r^ 00 



en •* i-o 00 o^ 00 LO 



I LO to r^ LO CT^ ^ t~^ 



OOOOLOOOO 

TJ- to "^ ■— I ■^ ■rt "^ 

I I I I I I I 



oc^ooooor^ 

tNvOCN'^tOLOLnvD 



lOLOCNOcOvO-^CN 
CN>— It— (i—tOOOO 



oooooooo 
oooooooo 

OOCNOOOOC 
LO-^tOlOtOtOtOtO 



Ooot^i~~Lnc7i>.orM 

CNl— It— It— IrHOOO 



+ 1.0 



+ .5 - 



> 
E 



I 

Q. 
O 

N 



Mg(0H)2 in tap water 



-.5 



10.5 1 1.0 Ili5 

PH 
Mg{OH)| gR in NSSC waste 



12.0 



-1.0*- 



Fig. 4.2 5 Zeta potential of magnesium solids in 

tap water and NSSC waste at varying pH 



87 




10.5 
Fig. 4.2 6 



Equilibrium concentrations of Mg"* and 
Mg{OH)'*" with Mg(0H)2 ai varying pH 



88 



greatest change in zeta potential with respect to pH occurred at pH 
11.1, which was close to 10.8, the pK for magnesium hydroxide. The 
positive charge on colloidal material formed in the presence of Mg 
in aqueous solution has been attributed by Loganathan and Maier (1975) 
to the formation of Mg(OH) . The equilibrium concentrations of 
Mg(0H3 and Mg are presented in Figure 4.26 from the equilibrium 
expressions given in Equations 4-3 and 4-4. The equilibrium concen- 
trations were equated to the activity of these species for these cal- 
culations. 

++ 2 -10.8 
Kgp = {Mg^^XOH"} = 10 C4-3) 

• + ++ -2.4 

K = {Mg(OH) }/{Mg }{0H} = 10 (4-4) 

The zeta potential as shown in Figure 4.25 increases with increasing 

+ ++ 
pH. The ratio of the singularly hydroxylated species Mg(OH) to Mg 

also increased with increasing pH. However, the equilibrium concen- 
tration of Mg(OH) decreased. As the concentration of Mg(OH) and 
Mg decreased, more magnesium hydroxide precipitated from solution. 
From the data in Figure 4.25 and the equilibrium graphs in Figure 
4.26, one can suggest the floe was positively charged in tap water 
due to its own nature and not due to the sorption of the singularly 
hydroxylated species. 

The zeta potential of the magnesium solids produced during the 
coagulation of the NSSC waste changed significantly to pH 10.7 and past 
pH 11.7. The first pH was approximately the minimum pH at which 90% 
color removal was obtained using magnesium coagulation. In the first 
portion of the zeta potential curve for the treated NSSC waste a 



89 



chemical compound was formed that v.as negatively charged. The col- 
loids maintained a relatively constant negative zeta potential 
through the pH range of maximiM color removal, pH 10.6 to 11.5. When 
the pH was raised to 12.5, the zeta potential increased to zero as a 
result of the formation of the neutral magnesium hydroxide. As pre- 
viously shown, color removal decreased past a coagulation pH of 11.5 
when NaOH was used to control pH. These data indicated as the pH 
exceeded 11.5, hydroxide ions probably displaced color bodies from 
the magnesium floe. The ability of the hydroxide ion to successfully 
compete for the reaction sites on the magnesium floe was probably 
reduced when Ca(OH) was used to control pH due to the complexing 
ability of the calcixrai ion. 

4-5 Magnesium Recovery and Recycle 

4-5.1 Recovery Methods 

Experiments were conducted to find a recovery process that would 
recover the magnesium in a usable form. One method of recovering the 
magnesixM was to acidify the magnesium sludge with CO2 gas. This 
would have returned the magnesium to solution but would have possi- 
bly returned the color to solution also. Taflin et al. (1975) ini- 
tially recovered magnesium from sludge by carbonation in Minneapolis. 
However, they had to discontinue magnesiiim recovery because of the 
color build-up in the recycled magnesium feed. 

One method of removing color from the magnesium was to inciner- 
ate the sludge before the magnesium was recycled. This would have 
oxidized the color bodies to CO2 and left magnesium in a usable 
form. The CO2 gas from incineration could be used to acidify the 



90 



incinerated magnesium sludge. 

4-5.2 Process Reversibility 

A color release experiment was conducted on a one liter sample 
of NSSC waste with an initial color of 2500. This waste was trea- 
ted with the optimum design doses of 150 mg/1 magnesium and 1000 
mg/1 Ca(OH) . The final pH of the coagulated waste was 11.0, and 
the final color was 200. The pH was lowered to 9.0, the color was 
measured, and then the pH was raised to 11.0 and the color was again 
measured. This oscillation of pH completely dissolved and reformed 
the sludge. When the 'sludge was dissolved, the color returned to 
2500. When the sludge was reformed, the color returned to 200. 
These data are presented in Figure 4.27. The reversibility presen- 
ted a problem for magnesium recovery by carbonation. Direct recycle 
with only carbonation would be feasible if the precipitated color 
remained on the CaCO, floe during carbonation. 

4-5.3 Color-Cation Interaction 

The carbonate alkalinity was low in the NSSC waste because it 
was prepared from low alkalinity tap water. An experiment was 
designed to determine if CaCO^ precipitation in situ with the 
colored waste would remove the NSSC color. If this happened, incin- 
eration was not required for magnesium recovery. If CO2 gas was 
used to recover the magnesium in actual plant operation, the return 
stream would contain a large amount of HCO^ alkalinity. The exact 
amount of HCOj-alkalinity would be controlled by the chemical 



91 



3000 - 



2000 



(0 



o 

o 
I 

a: 



2 1000 

o 

o 




Color = 2500 
Mg^ = 150 mg/l 
NaOH 



Color measured at pH 7.6 
after coagulated NSSC waste 
adjusted to pH 9 



^Color measured at pH 7.6 
after coagulated NSSC waste 
adjusted to pH It 



4 5 6 

pH cycles 



8 



Fig. 4.27 Color reversibility bar graph 



92 



interaction of magnesium, CO2, water and the organic acids present. 

In the CaCOj precipitation experiment, Na^CO^ and CaCOH)^ were 

dosed in equimolar amounts to produce CaCO^ sludge. The amount of 

CaCO- precipitated per liter ranged from 0.2 to 260 milliraoles. 

Since NaXO„ and Ca(OH)- were the source materials, the amount of 
2 3 Z 

CaCO, precipitated was determined by the calcium difference and 
checked by the carbonate difference before and after coagulation. 
The carbonate difference was measured by determining the total inor- • 
ganic carbon before and after coagulation. The residual color 
change as a function of the millimoles of CaCO^ precipitated is pre- 
sented in Figure 4.28. The dose data and the change in TOC concen- 
tration is summarized in Table 4-5. 

The color of the NSSC waste was reduced 65% when 265.5 milli- 
moles of CaCO, were precipitated from solution. This amount of cal- 
cium was not available from the Ca(0H)2 used to control the pH in 
the color removal process. The maximum amount of calcium available 
from the Ca(OH)^ dose for a color of 5000 would be 16.89 millimoles. 
If this amount of calcium was precipitated as CaCO^, there would be 
approximately a 4% color removal based on the data presented in 
Figure 4.28. 

It has been shown previously that the color removal process was 
reversible. The precipitation of CaCO^ did not remove a significant 
amount of color in this process. The color would still be present 
in the recovered magnesium solution. It was concluded that the color 
had to be removed from the magnesium floe by incineration in order 
to reuse the magnesium. 



93 



6000 



< Initial color increase due to chelation 




50 100 150 200 

CaC03 precipitated- mmol/l 
Fig. 4.28 Color remaining as a function of CoCO^ precipitation 



3000 



o. 2000 

c 

'c 

"5 

E 



■I 1000 
o 




5 .10 15 . 20 

MgF2 precipitated- mmol/l 

Fig. 4.29 Color remaining as a function of MgFg precipitation 



94 



eo 



< 






a: 

(/) o 

CO E- O 
2 WD O 

< ^ 



a 
< 



II 



o 

E- 



2 
O 

HH Q 

E- 



U 

m 
OS 

D, 

to 

O 
u 

U 



o 

o 
u 

en -J 

=) <; 

E E- 

O 1-1 

cd 2 

2 M 



O 
CO 
CM 

X 



+ 
+ 

O U 
•H ^-^ 



O 03 . — . 
S Oill to 
O 





-o 


tOT3 


0) 


O -H 


s 


U ^ 


f-l 


rt O 


o 


u w 


<4A 


>-< f-l 




ctf O 




C 'H 




•rH O 




tt. u 





T— I 

c o --- 

■r^ t^ bj:i 



C 

•H U — I 

MO --^ 

■ H [-■ M 

fH S 
O 



1 to > -H 

o o ^-, 



+ 


<n 




+ 


> 


<— 1 


Oj 


o 


*-^^ 


U 


6 


M 




<u 


Fi 




OS 




to 






o 


1-1 




u 


■*«N^ 




(N 


M 




rt 


e= 




2 







c3 



1— 1 

(M 




\0 








to 




o 


o 


O 


o 


o 


o 


o 


O 



CM 



\o 



to 

CM 



o 
to 






CM 



o 



o 


o 


O 


o 


o 


o 


o 


o 


LO 


o 


LO 


LO 


o 


to 


o 


o 


r-. 


to 


CJl 


r- 


vO 


\0 


LO 


l^^ 


LO 


vO 


■=t 


^ 


'a- 


to 


to 


1 — I 



LO 


o 


o 


LO 


o 


LO 


LO 


o 


to 


■o 


^ 


o 


\0 


o 


'^ 


o 


■=a- 


to 


CM 


o 


o 


.—1 


Ol 


CTl 



o o o 
o o o 

LO LO LO 



o 


o 


o 


o 


o 


o 


o 


o 


o 


o 


LO 


LO 


LO 


LO 


LO 



CM 



to 



LO 

to 



to 



CM 

00 



OO" 

to 
to 



•^ LO 

I--. LO 



00 



o 


o 


•^ 


t^ 


LO 


CO 


to 


I— 1 


•<+ 


to 


LO 


O 


CM 


'J- 


'^ 


MD 




r-^ 


CM 


LO i 


to 


\D 


to 


vO 










t— 1 


CM 


LO 


o 



1—1 
o 



00 



LO 

to 



o 



CM 

to 



o 

00 
LO 

to. 



o 



o 

CM 

to 



o 

CM 



LO 


o 


o 


o 


O 


o 


o 


o 


f~- 


LO 


o 


o 


o 


o 


o 


o 




CM 


LO 


o 


LO 


o 


o 


o 








rH 


CM 


LO 


o 
1—1 


o 

CM 



95 



A second color-cation experiment was conducted to further inves- 
tigate the degree o£ magnesium-color interaction. To investigate 
this, the amount o£ color removed by magnesium was determined for a 
magnesium precipitate that was not a hydrolysis product. MgF2 was 
selected as the precipitate because of its low solubility and pH 
independence. 

Different jars of NSSC waste with an initial color of 2950 were 
dosed with F from NaP and Mg from MgSO. . The Mg "''"'' dose was 175 
mg/1 for each jar. This was the same magnesium dose that would have 
been used for color removal by magnesium coagulation at high pH. 
All jars were in a state of MgF2 supersaturation. The general defi- 
nition of the supersaturation ratio (S) is S = (Q/K) , where Q is 
the ion product, K is the equilibrium constant, and n is the number 
of ions in the neutral molecule (n = 3 for MgF^) . The jars were 
allowed to stand for 24 hours before samples were taken. The floe 
formed was very small and not nearly as voluminous as the magnesium 
floe produced at high pH. The residual color as a function of the 
millimoles/liter of MgF2 precipitated is presented in Figure 4.29. 

The initial color of the NSSC waste was reduced 55% when the 
initial MgF2 supersaturation ratio was approximately 16. At this 
point all of the magnesium dose, 7.3 mmole, had been precipitated 
from solution. In the MgF^ experiment, the precipitation of 1 mmole 
of magnesium removed 231 Pt-Co color units. If more magnesium had 
been used, the amount of color per mmole of magnesium precipitated 
might have been larger. However, a significant magnesium-color inter- 
action had been demonstrated in a non pH dependent chemical reaction 
involving magnesium precipitation. 



96 



The data in Table 4-5 shows that the organic carbon level was 
reduced 40% and the color was reduced 66% when 266 iranole/1 of CaCO_ 
were precipitated. The molar ratio of Ca removed to CO, removed 
was consistently less than one as shown in Table 4-5. This indicated 

that the color removal mechanism with Ca precipitation was similar 

++ 
to Mg precipitation. The negatively charged color body acted as a 

contributing anion in a chemical reaction. When CaCO, was precipi- 
tated from a NSSC waste, the removal of one mmole of calcium reduced 
the color 12.4 Pt-Co units. 

In the lime-magnesium color removal process the precipitation 
of one mmole of magnesium removed approximately 360 color units from 
solution. Comparison of the ratios of color removed per mmole of 
magnesium precipitated demonstrates that magnesium is more color reac- 
tive than calcium. The magnesium sludge produced at high pH was more 
effective than the sludge produced in the non pH dependent MgF^ 
experiment. The larger floe produced by magnesium coagulation at 
high pH possibly provided more sites for magnesium-color intei-action. 
From these data it was concluded that magnesium was responsible for 
color removal in the lime-magnesiiom color removal process. 

4-5.4 Chemical Equilibrium of Mg •-CO^-H 

The chemical equilibrium of the Mg -CO^-HnO system was described 
by constructing an activity diagram, a solubility diagram and a CO2 
predominance diagram. These diagrams are presented in Figures 4.30, 
4.31 and 4.32. The equations used to construct these diagrams are 
presented in Table 4-6. A predominance diagram was constructed for 



97 




8 
PH 



10 



II 



Fig. 4.30 Activity ratio diagram for iogCj=-l. 
Equations defining relative activities 
ore given in Table 4.6. ForOMglOH)^, 
(DMgC03, @MgC03- SHgO, and 
(DMg4(C03)3(0H)2-3H20 which are 
all solid forms. 



98 



o 

o ^ 




■H- . 

Fig. 4.31 Solubility diagram of Mg in a 

Cj= iO~ M carbonate system. Equa- 
tions defining relative activities for 
all species shown are given in table 4.6. 
MgCO^SH^O, Mg^(C02)2(OH)2-3H20 
and MglOH) are solid forms 



99 





\ ' 


-1 1 1 1 1 





■ \^ 


MgCOj- 3H2O 


cv 2 


- 


\ 


O 




\ 


o 


■H- 


\ 


a. 4 


\ 


o 
-J 
1 


Mg 


\Mg4(C03)2(0H)2-3H20 


6 




- 


8 


1 1 


_i 1 


Mg{0H)2 



7 8 

PH 



10 



■H- 



Fig. 4.32 Predominance diagram for log Mg 

MgC03-3H20 is only stable thermo- 
dynomicQlly at high Pqq . MgCO -SH 0, 
Mg^{C02)2(0H)2-3H20 apd MglOH)^ 
are solid forms 



TABLE 4-6 



CHEMICAL REACTION AND pK VALUES CONSIDERED 
FOR Mg'-'-C02-H20 SYSTEM 



HCO3" = H"" + CO3 



Mg4(C03)3(0H)2-3H20(s) = 4Mg*'' + 3CO3 + 20H" + 5H2O 



2' 



100 



0, 



„ ^. „ pK at 25 C 

Reaction ^ 



H20=H^-OH- 14.0 a 

H,CO, = 2H" + Co! 16.6 a 
2 3 aq ^ 

10.3 a 



H2CO3* = H"^ + HCO3- 6-2 ^ 



1.47 b 
29.5 b 



4v.--3^3v 

MgC03-3H20(s) = Mg""* + CO3 + 3H2O 5.4 b 

++ ^^= 4 9 b 

MgC03(s) = Mg + CO3 

Mg(OH)„(s) = Mg"^ + 20H' 10-85 "" 



a Meites, L. , Handbook of Analytical Chemistry , McGraw-Hill, 
N.Y., N.Y., (1962) 

b Stumm, W., Morgan, J.J., Aquatic Chemistry , John Wiley § Sons, 
N.Y., N.Y., (1970) 

c Day, R.A., Underwood, A.L., Quantitative Analysis , Prentice HalL 
Englewood Cliffs, N.J., (1967) 



101 



a Mg** concentration of lO""-^ M. The activity ratio" and solubility 
diagrams were constructed for a total alkalinity o£ 10 M. 

The stable species were determined as a function of pH from the 
activity ratio and solubility diagrams in Figures 4.30 and 4.31. The 
activity ratio diagram was drawn using Mg"*"* as the reference state. 
MgCOH)»(s) was the stable species above pH 11.2. From pH 11.2 to pH 
8.5 the controlling species was Mg^ (00^)3 (OH) ^ * 3H2O (s) . Below pH 8.5, 
MgCOj-SH 0(s) was the stable species. These pH ranges of the stable 
species will shift with variations in the total alkalinity. The 
practical significance of these diagrams was to identify the species 
controlling the solubility of Mg"^"^. For a Cj of 10"-^ M, this species 
is MgC0^-3H20(s) for an open system using CO^ gas as the proton source. 

The dominance of MgCO^-SH^OCs) is seen in the solubility diagram 
in Figure 4.31. For a C-p of 10" M an equilibrium point was reached 
at pH 7.2. This is identified by the intersection of the equilibrium 
lines in the activity diagram of HCO3" and MgC03-3H20(s) . The bicar- 
bonate ions from the solubilizing of CO2 gas are in equilibrium with 
the solid MgCO •3H20(s) . The solid was dissolved by reaction with a 
proton from H2CO3 . This was represented by the following reaction 
and pH calculations. 

MgC03'3H20(s) + H"" = Mg""" + HCO3" + 3H2O (4-5) 

K = 10"''^ -^^ = (Mg"""") (HC03~) (4-6) 

(H*) 

(HCO3") = a^C^ (-4_7-, 

at point of maximum solubility 
2(Mg++) = (HCO3-) 



102 



104-9 = (a^C^)2/2(H"') " (4-8) 

pH = 7.2 (4-9) 

It was interesting to note from the activity ratio diagram in Figure 
4.30 that the species MgCO,(s) never does exist at equilibrium. 

A predominance diagram is presented in Figure 4.32. In the pre- 
dominance diagram the species in equilibrium with CO2 gas is shown. 
For the CO2 concentration in flue gas (10%-14% CO2) the equilibrium- 
species was MgCO„-3H20(s) . 

The magnesium concentration in the recovery tank depends on the 
pressure of CO2 gas and the equilibrium species, MgC02'3H20(s) . For 
a 10% CO2 gaseous mixture in an aqueous solution, 3.89 x 10 mols of 
CO2 are dissolved at 20°C. A solution containing solid Mg(0H)2(s) 

mixed with a 10% C0„ gas would come to equilibrium at pH 7.5, The 

-2 
total alkalinity calculated at equilibrium was 6.0 x 10 M, with 

94% of it in the HCO-~ form. The maximum calculated magnesiiora solu- 
bility considering MgC02'3H20(s) as the dominant species was 863 mg/1. 

The laboratory system employed was never at true equilibrium. 
The source of the magnesium was incinerated MgO(s) or Mg(0H)2(s). If 
Mg dissolution from Mg(OH)„(s) was kinetically favored as compared 
to MgC0„*3H20(s) formation, a supersaturated system with respect to 
MgC0_-3H20(s) could exist. The presence of organic acids in the 
recarbonation basin would increase the Mg solubility, and the spe- 
cies controlling Mg solubility may be an organo-metallic compound. 
This would be more likely when incineration was not used to remove 
color, because the organics would not have been converted to CO . 



103 



4-5.5 Sludge Incineration 

The colored sludge that was to be incinerated was prepared by 
treating two separate 40 liter volumes o£ NSSC waste with 350 mg/1 
Mg** and 1250 mg/1 of Ca(0H)2. An average of four separate analyses 
of NSSC waste and sludge is presented in Table 4-7. All sludge 
samples were dried at 103°C until constant weights were obtained 
before the sludge was incinerated. 

The magnesium dose plus the magnesium in the NSSC waste was 362 
mg/1. The magnesium concentration of the supernatant was 14 mg/1. 
1082 mg of solids resulted from the incineration of the sludge pro- 
duced in one liter of waste. The incinerated solids contained 32.2% 
magnesium, A mass balance on magnesium revealed that 362 mg were 
available before color removal and 362 mg of magnesium were accounted 
for after color removal. The total recovered magnesium in the solids 
was 96%. Some magnesium was lost in the supernatant. This was pro- 
bably due to the filtering of the sludge through a no. 40 Whatman 
ashless filter which passed some very small particles of solid magne- 
sium salts. These would have been retained by the 0.80 micron Milli- 
pore filtering apparatus used in the color determination. 

A chemical representation of the incineration of the colored 
sludge is as follows: 

Mg, (OH), _R = NfeO + COo + other gases (4-10) 
11.5 ■^ 

where R is the symbol for the color bodies 
Since the color bodies would be oxidized to CO2 and other gases 
by incineration, an investigation of the removal rate of color by 
incineration at varying times and temperatures was implemented. The 



104 



TABLE 4-7 

AVERAGE CHARACTERISTICS OF A SLUDGE PREPARED 

BY COAGULATING A NSSC WASTE AT pH 11 

WITH 350 mg/1 OF Mg"""" AND 1250 mg/1 OF CaC0H)2 

INITIAL COLOR OF NSSC WASTE = 4925 



Parameter Value 



Suspended solids 1800 mg/1 

Nonvolatile suspended solids ' 1082 rag/1 
Volatile suspended solids 718 mg/1 

% Magnesium in NVS 32.2% 

SVI 352 



105 



primary intent was to find the minimum amount of time and minimum 
temperature required to remove the color bodies from the magnesium 
solids. However, a limiting factor influencing the recovery of the 
magnesium would be the chemical species formed during incineration. 
The specific gravity of magnesium varied considerably with the tem- 
perature and. time of incineration. This is presented in Table 4-8. 
A decrease in reactivity of MgO was paralleled by an increase in the 
density of MgO resulting from increasing calcination temperatures. 
According to Harper (1967) , the freshly formed MgO had a high surface 
area. This area was reduced as heating temperatures were increased. 
The porosity of the oxide was reduced until, at a sufficiently high 
temperature, dead-burned magnesia resulted. The dead-burned magnesia 
resulted from compounds incinerated at temperatures in excess of 
900°C and was very unreactive. Harper (1967) also found MgO prepared 
in the range of 400-900. C, called caustic burned magnesia, was readily 
soluble in acid and rapidly hydrated in cold water. 

Standard Methods (1971) reports that it has been found that waste- 
water and effluent residues usually obtain constant weight after 15-20 
minutes of ignition at 550 C. It was decided to ignite the sludge 
samples beginning at a temperature of 150 C and progressing to a final 
temperature of 850°C in increments of 100 C. Magnesia incinerated at 
temperatures more than 900°C would not have been soluble in a carbo- 
nated solution. Three different time increments were investigated at 
15, 30 and 60 minutes. The 30 and 60 minute tests were discontinued 
for temperatures past 550 C. This was because the incineration of 
color was complete after 15 minutes at the high temperatures. 



106 



TABLE 4-8 
MgO REACTIVITY AS AFFECTED BY TEMPERATURE 



Temperature °C Specific Gravity 



600 2.94 

700 3.04 

850 3.22 

1000 3.39 

1200 3.48 

1400 3.52 

1500 3.56 



Source : Kirk , J . E . , Othmer , D . F . , Encyclopedia of Chemical Technology , 
Second Edition, 12, 'John Wiley and Sons, N.Y., 1967, 



107 



The results o£ this experiment are presented in Table 4-9. 
Since the object of incineration was to remove color bodies in order 
to obtain reusable magnesium, a color/Mg ratio was calculated and 
is presented in Figure 4.33 as a function of incineration time and 
temperature. The incinerated sludge was resolubilized with hydro- 
chloric acid. The color and magnesium concentration in the resolu- 
bilized sludge were then determined. These two parameters were divi- 
ded to give a color/Mg ratio. A small ratio (less than one) would 
indicate that the magnesium was in a reusable form for color removal. 
Figure 4.33 shows the color/Mg ratio was minimized at 550 C for all 
incineration times tested. A ratio of 0.25 meant that over 98% of 
the color on the magnesium solids was removed by incineration and 
the magnesiiim could be recycled without a significant color concen- 
tration in the recycle stream. 

4-5.6 Magnesium Recovery 

Magnesium recovery by recarbonation was selected as the recovery 
method because of the availability of free CO^ in flue gas, which 
usually is available at pulp and paper plants. The cost of recover- 
ing the magnesium with H2SO4 was $0.24/1000 gal. It was concluded, 
therefore, that H-SO. was too expensive to use as a means of recovery. 

In the laboratory study samples of nonvolatile solids were slur- 
ried in 200 ml of deionized water in a 400 ml beaker. The beaker was 
placed on top of a magnetic stirrer and the slurry was agitated by 
both the stirring bar and a 10% CO2-90% air gaseous mixture. The 



108 







TABLE 4-9 










INCINERATED SOLIDS ANALYSIS 






Incineration 
Temperature 


Incineration 
Time 
Minutes 


% NVS 


% VS 


% Magnesium 
of NVS 


Color 
Pt-Co 


Cone. 

NVS 

mg/1 


150 


15 

30. 

60 


99.12 

97.51 

■ 99.02 


0.88 
2.49 
0.98 




15.4 
15.7 
15.5 


1157 
1419 
1325 


378.8 
379.2 
390.0 


250 


15 
30 
60 


92.72 
90.77 
87.06 


12.94 
9.23 
7.28 


16.5 
16.9 
17.6 


1118 
793 
368 


320.4 
280.0 
296.3 


350 


15 
30 
60 


85.17 
85.64 
81.22 


14.83 
14.36 
18.78 


18.0 
17.9 
18.8 ■ 


402 

357.4 

303.5 


309.4 
285.1 
209.3 


450 


15 
30 
60 


62.59 
59.12 
57.24 


37.41 
40.88 
42.76 


24.4 
25.9 
26.70 


197.0 
39.3 
26.2 


61.4 
244.3 
282.7 


550 


15 
30 
60 


54.94 
53.28 
55.31 


45.06 
46.72 
44.69 


27.80 
28.70 
27.70 


13.1 
21.9 
11.0 


264.3 
235.3 
259.6 


650 


15 


50.72 


49.28 


25.90 


24.05 


223.8 


750 


15 


47.43 


52.57 


36.67 


41.52 


705.25 


850 


15 


42.13 


57.87 


39.70 


48.07 


307.3 



109 





AT 


o 


15 min. 


• 


30 min. 


D 


60 min. 


at= 


Incineration 




time 




350° 450** 550'' 650* 

Incineration temperature {°C) 
F\q. 4.33 Coior/Mg'^ ratio as a function of incineration temperature 



110 



gaseous mixture was introduced into the slurry through a small porous 
stone diffuser. The 10% CO -90% air gaseous mixture was controlled 
by rotameters connected to a 100% CO tank and a laboratory air source. 
The total flow of the gaseous mixture was 2122 ml/min. This was the 
maximum flow rate that could have been implemented without losing 
some of the slurry due to turbulance. The pH of the slurry was con- 
tinually monitored for the total recarbonation period. The tempera- 
ture of the slurry was approximately 20°C. 

As illustrated in section 4-5.4, the amount of magnesium resolu- 
bilized from the incinerated solids by carbonation at equilibrium was 
controlled by the concentration of 'CO2 gas, not the amount of CO gas. 
However, the amount of CO^ gas could affect the rate of the dissolu- 
tion reaction due to surface area contact. A decrease in the tempera- 
ture of the. carbonation reaction or a decrease in the CO pressure 
would decrease the soluble magnesium. This can be seen from the ther- 
modynamic expression for the equilibrium constant and the chemical 
expression of activities for the equilibrium constant. 



log K = - AG° 



2.303 RT 



(4-11) 



Because log K for MgC02"3H20(s) is negative for the recovery 
reaction, a decrease in temperature would decrease the amount of mag- 
nesium in solution at equilibrium. An increase in the CO pressure 

would increase the CO activity which would increase the amount of 

++ 
Mg and HCO^- at equilibrium. The temperature effects on the rate 

of MgCO^-SH^OCs) precipitation are illustrated in Figure 4.34 in data 



Ill 




20 



80 



100 



40 60 

Time - minutes 
Fig. 4.34 Precipitation of IV1gC03-3H20 by aeration at various temperatures. 
Source ' Blaci<,A.P, EPA #12120 HMZ, Sept., 1 974. "Full Scale Studies of the 
Magnesium Carbonate Water Treatment Process at Montgomery, Alabama and 
Melbourne, Florida. 



112 



gathered by Black and coworkers in Dayton, Ohio and Montgomery, 
Alabama. The effect of increasing CO2 pressure is illustrated in 
the predominance diagram presented earlier in Figure 4.32. The mag- 
nesium recovery data are presented in Table 4-10. In each recovery 
experiment, the time of carbonation was 120 minutes. The pH readings 
and magnesium concentrations were measured at the time intervals of 
0, 5, 15, 30, 45, 60, 90 and 120 minutes. The nonvolatile solids 
content of the slurry was varied from 5318 mg/1 to 119,692 mg/1. 

The plot of pH verses carbonation time in Figure 4.35 indicates 

that an equilibrium pH was approached in the laboratory studies. The 

++ 
actual pH was close to that predicted by the Mg -CO2-H2O system m 

equilibrium wit}i a 10% CO2 gaseous stream. In this system the solid 

formed at equilibrium was MgC0„-3H20fs) . The theoretically predicted 

equilibrium pH was 7.6 after a carbonation time of 120 minutes for 

the laboratory studies. The impurities present and the lack of a 

true equilibrium condition could have accounted for the difference 

in pH values. The agreement between the equilibriiMi pH and the actual 

pH supports the formation of a magnesium carbonate compound in the 

recovery process. 

In Figure 4.36 the magnesium concentration after the carbonation 
of nonvolatile solids is shown as a function of .carbonation time. 
Figure 4.37 presents the degree of magnesitmi recovery during carbo- 
nation as a function of the nonvolatile solids concentration. 

The data in Figure 4.36 indicate that if the nonvolatile solids 
concentration was increased, more magnesium was recovered during car- 
bonation. However, the degree of the magnesium recovered from the 



113 



TABLE 4-10 

CARBONATION OF INCINERATED SLUDGE AT VARYING CONCENTRATIONS 
OF NONVOLATILE SOLIDS FOR MAGNESIUM RECOVERY 



Carbonation 
Test 



Nonvolatile 

Solids mg/1 5,318 22,950 33,561 74,550 119,692 

Total Available 

Mg"^"^ mg/1 1,300 5,000 7,317 16,252 26,095 

++ 
Final Mg 

Concentration 

mg/l 1,300 4,803 5,724 9,425 12,776 

Final Color 

Pt-Co 234 2,480 4,368 7,862 6,552 

Color/mg 

Mg""* Ratio 0.18 ' 0.56 . 0,60 0.48 0.25 



114 



Mg"" mg/l 

•300 

5000 

7317 
16,252 
26,095 




25 SO 75 100 

Carbonation time (minutes) 
Fig. 4.35 pH as a function of carbonation at various 
nonvolatile solids concentrations 



125 



115 



14,000 



12,000 



E 



■o 10,000 






c 

o 



c 
«> 
o 



8000 



6000 



E 
■Jo 

§> 4000 
o 



2000 



NVS-mg/ Mg-mg/l 



• 


5318 


1300 


■ 


22,950 


5000 





33,561 


7317 


n 


74,550 


16,252 


▲ 


1 1 9,692 


26,095 




50 75 

Carbonation time (minutes) 
Fig. 4.36 Magnesium recovered as a function of carbonation time 



116 



Time- minutes 
o 120 




2.5 m 7.5 lao 

Non-volatile solids mg/l X 10""* 

Fig. 4.37 % Magnesium recovery as a function of nonvolatile 
solids concentration 



12.5 



117 



nonvolatile solids was reduced as the nonvolatile solids concentra- 
tion was increased. After 120 minutes of carbonation, the soluble 
magnesium concentration increased from 1300 to 12,776 mg/1 when the 
nonvolatile solids concentration was increased from 5318 to 119,692 
mg/1. The degree of magnesium recovery decreased from 100 to 49 per 
cent for these two experiments. 

The magnesium concentration after carbonation was never limited 
by the formation of a solid magnesium compound for the carbonation 
times tested. This would have been apparent if the magnesium concen- 
tration had become constant during carbonation and some magnesivun 
still remained in the nonvolatile solids. As shown in Figure 4.36, 
the magnesium concentration was always increasing during the carbo- 
nation process if there was any magnesium remaining in the nonvolatile 

solids. 

The maximum theoretical magnesiiim concentration at equilibrium 
is 866 mg/1 if MgC02-3H20(s) was the controlling solid phase. This 
was the controlling solid phase predicted by the theoretical 
Mg*"'-C02-H20 system presented in Figure 4.32. The controlling solid 
phase was not determined by these experiments. It might have been 
possible for the system to become supersaturated with respect to 
MgCO,'3H^OCs) since the solid magnesium compound used for carbonation 
was not MgC02-3H20(s) . Supersaturation could occur if the dissolution 
rate of MgO(s) or MgC0H)2(s) was kinetically favored with respect to 
niicleation of MgCO^-SH 0(s) . Another possibility was that other 
species in the incinerated solids were involved in the controlling 
solid phase and equilibrium was not achieved during carbonation. 



118 



4-5.7 Magnesium Reuse 

Incineration and carbonation processes were used to recover the 
magnesium used in the color removal process. Each of these processes 
affected the chemical form of the magnesium. A magnesium reuse experi- 
ment was performed in order to determine if the recovered magnesium 
could be successfully reused in the coagulation process. 

The magnesium was recycled twice. Following each use, the mag- 
nesium was recovered by incineration at 550°C for 15 minutes and then 
carbonated with a 10% C0„ gas. The nonvolatile solids concentration 
was 5318 mg/1 during carbonation. . Some of the nonvolatile solids 
were not dissolved during carbonation. These remaining solids were 
recycled with each magnesium reuse. The magnesium dose was based 
on the soluble magnesium in the carbonated liquor and the coagulant 
dose relationship shown in Figure 4.8. The data from these magne- 
sium reuse experiments are summarized in Table 4-11. 

After incineration the sludge remaining after the first use of 
the magnesium contained 99.6% of the initial magnesium. However, 
only 66% of the magnesium was in a soluble form after the sludge 
from the first use was carbonated. The color of this solution was 
600. The remaining magnesium was still in the nonvolatile solids 
that were not dissolved during carbonation. 

When the sludge from the second magnesium use was carbonated, 
92% of the magnesium was in a soluble form. The color of the reco- 
vered liquor was 80. The difference between the per cent of magne- 
sium- solubilized after the first and second carbonation processes 



119 



TABLE 4-11 . 

COLOR REMOVAL BY LIME -MAGNESIUM COAGULATION .USING THE SAME MAGNESIUM 
THREE TIMES. MAGNESIUM RECOVERY WAS ACCOMPLISHED BY INCINERATION 

AND CARBONATION FOLLOWING COAGULATION. NONVOLATILE SOLIDS NOT DIS- 
SOLVED BY CARBONATION WERE RECYCLED WITH THE RECOVERED MAGNESIUM. 



Use of Per cent o£ Per cent o£ 

Magnesium Original Magnesium Color Removed 

In Soluble Form 



First 100 92 

Second '66 -92 

Third 92 91 



120 



was probably due to the incomplete combustion o£ the color bodies 
after the first magnesium use. The incinerated sludge contained 
77% nonvolatile solids after the first magnesium use. The average 
nonvolatile solids reduction in the previous experiments reported 
in Table 4-9 for 550°C and 15 minutes was 55%. The solids loading 
rate for the incineration process was never optimized. 

Practically all (99.6%) of the magnesium had been recovered 
after it had been used in the color removal process. Greater than 
90% of the initial color was removed with each of the three magne- 
sium uses. The magnesium dose was the same for each use. The 
recovered magnesium and the undissolved solids removed as much color 
in the second ard third uses as did the fresh magnesium in the first 
use. It was concluded that the magnesium could be successfully 
recovered and recycled in the lime -magnesium color removal process 
after the incineration and recovery processes. 

A second set of magnesiiim reuse experiments was conducted to 
determine the effectiveness of the recycled solids that were not 
dissolved in the carbonation process. The recovered magnesixim solu- 
tion was filtered through a 0.80 micron Millipore filter to remove 
the solids. The per cent of color removed by the used magnesium 
with no solids was compared to the per cent of color removed by an 
equivalent amount of unused magnesium. The data from this experiment 
are presented in Table 4-12. 

The reused magnesium that contained no suspended solids removed 
13% less color than did an equivalent amount of an unused magnesium 
when both were used separately in the color removal process. The 



121 



TABLE 4-12 

COLOR REMOVAL BY LIME-MAGNESIUM COAGULATION USING THE SAME MAGNESIUM 
TWICE. MAGNESIUM RECOVERY WAS ACCOMPLISHED BY INCINERATION, CARONA- 
TION AND FILTRATION FOLLOWING COAGULATION. NO SUSPENDED SOLIDS WERE 
RECYCLED WITH THE RECOVERED MAGNESIUM. INITIAL COLOR = 5000. 



Use of 


Magnesium 


Final 


% 


Color 


Magnesium 


Dose - mg/1 


Color 


Reduction 


First 


300 


550 




8d 


Second 


300 


1190 




76 


Second 


343 


1096 




78 


Second 


395 


943 




81 



122 



presence of the incinerated solids was a significant aid to the 
reused magnesium in the color removal process. These solids prob- 
ably provided nucleating surfaces for the forming solids phase. 



CHAPTER 5 



DESIGN OF A COLOR REMOVAL PROCESS FOR A NSSC 
WASTE USING MAGNESIUM COAGULATION AND RECOVERY 

The color removal process consists o£ several different unit 
operations. These are coagulation, sedimentat^ion, vacuum filtration, 
incineration and carbonation. The design of each of these unit opera- 
tions was considered separately in this chapter. The design para- 
meters were based on the treatment of a NSSC waste with an initial 
color of 5000 and a flow of 1 mgd. 

5-1 Coagulation 
A velocity gradient of 1000 sec"! was recommended by the AWWA 
(1969) to achieve adequate coagulant dispersion during the rapid mix 
process in a contact time of 20 seconds. The tank volume and energy 
requirements were determined from Equation 5-1. 



G = 



h 
550 P 



Vu 



(5-1) 



where: G = velocity gradient - sec 

P = water horsepower - hp 

3 
V = tank volume - ft 

2 • 
u = viscosity - lb sec/ft 

The tank volume and energy required were 30 ft^ and 1.5 horsepower. 



123 



124 



The amount of magnesium and CaCOH)™ to treat the waste was deter- 
mined from the optimum dose equations presented earlier in Figures 4.7 
and 4.8. The CaC0H)2 requirement was 10,425 Ibs/mgal, and the magne- 
sium requirement was 2502 Ibs/mgal. 

S-2 Sedimentation 
The information required to develop basic design criteria for 
a secondary sedimentation tank was obtained from the batch settling 
tests and Equations 5-2, 5-3 and 5-4. 

A^ = Q/Vj^ C5-2) 

2 

A = Area required for clarification - ft 

V,- = Hindered settling velocity - fpm 
Q = Flow cfm 

At = Qt^^/Ho (5-3) 

2 
A. = Area required for thickening - ft 

t^^ = Time required to reach desired concentration - min 

Hq = Initial height of sludge interface - ft 

HqCo = HuCu C5-4) 

Cq = Initial solids concentration - mg/1 
Hu = Final height of sludge interface - ft 
Cy = Desired solids concentration - mg/1 



125 



The settling velocity was determined to be 0.187 £t/rain from 
the hindered portion of the settling curve in Figure 5.1. The ini- 
tial and final heights of the sludge interface were 1.12 ft and 0.20 
ft. Equation 5-2 gave 610 ft^ as the area required for clarification. 
The area required for thickening the sludge to a final concentration 
of 11,520 mg/1 was found from Equations 5-3 and 5-4 to be 1020 ft^. 

The area required for thickening the sludge to a final concen- 
tration of 8000 mg/1 was larger than the area required for clarifi- 
cation and therefore controlled the area for settling. A design area 
of 1020 ft would provide adequate. settling area for the treatment of 
any NSSC waste with an initial color of 5000 or less. This was 
because the settling area was determined by the amount of solids pro- 
duced during treatment and a smaller amount of solids would be pro- 
duced from the treatment of a weaker waste. 

The solids loading that would be passed onto the vacuum filter 
would be 15,012 Ibs/mgal. If carbonation were used to recover the 
magnesium, the calcium present in the Ca(0H)2 would precipitate as 
CaC02 during coagulation. This would increase the mass of solids 
passed onto the vacuum filter. Since 1250 mg/1 Ca(OH3 2 was used to 
adjust the waste to pH 11, 16.9 mM of Ca** was available to precipi- 
tate as CaCO^. The magnesium concentration in the recycle stream 
was 1300 mg/1. To achieve the required magnesium dose of 300 rag/1, 
the recirculation ratio of the waste stream to the recycle stream 
would have to equal 0.23. The total carbonate concentration in the 
recycle stream was not determined. For design purposes, all of the 
calcium from the Ca(0H)2 dose was assumed to precipitate as CaCO,. 



126 



JO 

E 



.^ 600 



S 500 



C 







Initial Concentrations 

Sus. solids = 1800 nng/l 

Color = 5000 

Mg*= 350 mg/l 

Lime= i?.50 mg/l 

Polymer = 3.0 mg/l Hydrolyzed 

Polyacryiamide 
(Anionic) 



10 20 

Settling time minutes 



Fig. 5.1 Design data for sedimentation 



127 



This assumption is theoretically valid because electroneutrality 
requires that two moles of HCO," are available for every one mole 
of Mg'^'*' solubilized in the recovery process. Approximately 107 mM 
of HCO^" would be available in the recycle stream. In the coagula- 
tion tank at pH 11, this would be converted to CO3 and precipitate 
approximately all of the calcium as CaCO . Approximately 16.9 mM of 
CaCO, or 1690 mg/1 would be added to the solids passed onto the 
vacuum filter. Because the CaCOj formed acts as a settling aid, no 
additional allowances were made in the settling area calculations. 
The solids loading data are summarized in Table 5-1. 

The Ca(OH32 dose in Table 5-1 is the design dose determined for 
color removal from a NSSC waste by magnesium coagulation. The solids 
content of the sludge without the CaCO- was determined in Chapter 4. 
The total solids represented the sum of the solids from the precipi- 
tation of CaCO,. 

The volume of sludge coming from the thickening operation was 
determined for a color of 5000, and was 175 ml per liter of waste 
treated. The overflow rate from the sedimentation basin would be 
1034 gpd/ft^. The per cent solids in the settled sludge including 
the CaCO- precipitate would be 1.9.9%. . 

5-3 Vacuum Filtration 
Vacuum filtration studies on magnesium sludge were performed 
at Melbourne, Florida by Black (1974) . Sludge was produced by trea- 
ting a surface water source for color removal. The solids content 
in the sludge after vacuum filtration was 45%. In those studies. 



128 



TABLE 5-1 
SOLIDS LOADING FROM SETTLING BASIN 



Color Pt-Co 5000 

Lime dose mg/1 1250 

Ca"*"*" mM from lime 16.9 

CaC03 mM precipitated 16.9 

CaCOj mg/1 1690 

Solids mg/1 1800 

Total solids mg/1 3490 

Ibs/mgal 29,107 



129 



2 
filtration rates ranged from 11.7 to 20.0 lbs/ft /hr. The sludge 

produced after NSSC waste treatment would have a lower CaCO^ con- 
centration than the Melbourne sludge, thereby producing a lower fil- 
tration rate, Liptak (1974) found that compounds precipitated by 

2 
Ca(0H)2 can be filtered at a rate of 2 to 6 lbs/ft /hr on a rotary 

- • 

vacuum filter and the solids content in such a filter cake would 
vary from 20 to 30%. A design rate of 6 Ibs/ft^/hr and 20% solids 
content was selected for vacuum filtration of the NSSC sludge 
following coagulation. « 

Based on a design rate of 6 lbs/ft /hr, the total amount of 
solids processed would be 144 lbs/ ft /day. For each million gallons 

of waste treated, 29,107 lbs of solid has to be vacuum filtered. The 

2 
total area required for vacuum filtration was 202 ft . 

A second vacuum filter would be required if recalcination was 

used to recover the Ca(0H)2. The solids filtered would be the CaCO^ 

solids produced in the coagulation basin from the reaction between 

Ca(0H)2 and the carbonated recycled stream. These solids would 

amount to 14,094 Ibs/mgal. A design rate of 12 Ibs/ft^/hr was used 

to determine the size of the vacuum filter for the CaCO, solids. The 

2 
area required to remove the CaCO^ before lime recovery was 49 ft . 



5-4 Incineration 
To estimate the energy requirement of the incineration process, 
it was necessary to consider both the sensible and latent heat 
requirements of the NSSC sludge. Lignin sludge has a 8,000 to 10,000 
Btu/lb fuel value and contains 70% volatile solids. The fuel value 



130 



of the NSSC sludge was estimated at 5140 Btu/lb based on a 45% 
volatile solids content in the sludge. 

The specific heat of the solids was estimated to be 0.75 
Btu/lb °F from Liptak (1974) . The sensible heat requirement was 
determined to be 18,540,516 Btu from the specific heat, material 
weight and 550°C incineration temperature of the sludge. A latent 
heat requirement of 139,483,500 Btu was determined from the amount 
of v/ater that had to be evaporated during the incineration process. 
The amount of available energy was determined from the fuel value 
of the sludge to be 147,151,500 Btu. The total heat requirement 
exceeded the available heat by 7,688,000 Btu. This heat would have 
to be supplied ^ly use of a fuel such as oil or natural gas. 

5-5 Carbonation 

The operating conditions of the carbonation process were selec- 
ted from the recovery experiments described in Chapter 4. These 
operating conditions could be further optimized to increase effi- 
ciency. In the recovery experiments, 100% magnesium recovery was 
only obtained after 0.957 grams of incinerated solids was carbonated 
in 180 mis for 45 minutes with a 10% C02-90% air gaseous mixture 
flowing at 2122 ml/min. 

The solids from the recovery experiments were increased by the 
theoretical amount of CaCO_ that would be added to the sludge when 
carbonation would be used to recover the magnesium. The additional 
CaCO_ would probably not affect the amount of magnesium recovered 
since it would not enter into the Mg"^'^-H20-C02 equilibrium system. 



131 



The following design parameters were developed for a 1 mgd 
flow into the color removal process using the operating conditions 
that were previously selected. The daily total solids including 
the additional CaCO, input to the carbonation tank would be 22,351 
lbs. Since a recirculation ratio of 0.23 was determined in section 
5-2, 230,000 gallons of carbonated recovery liquor would be pro- 
duced per day. The recovery process required approximately 161,435 
lbs of CO2 per day. The most economical source of CO2 gas is flue 
gas. The flow rate of the flue gas into the carbonation tank would 
be 9167 cfm if the flue gas contained 10% CO^. A design summary is 
presented for the lime-magnesium color removal process in Table 5-2, 
and a flow chart of this process is presented in Figure 5.2. 



132 



TABLE 5-2 

DESIGN SUMMARY FOR THE TREATMENT OF A NSSC WASTE WITH AN INITIAL COLOR 
OF 5000 AND A FLOW OF 1 MGD BY THE LIME -MAGNESIUM COLOR REMOVAL PROCESS 



Unit Process 



Design Parameter 



Comments 



Rapid Mix 



20 sec 



G = 1000 sec 



-1 



Coagulation and 
Sedimentation 



10,425 lbs Ca(0H)2 
2502 lbs Mg"*"* 
25 lbs 835A 



1020 ft^ 

29,107 lbs solids 



5% hydrolyzed poly- 
acrylamide polymer 

Sludge thickened 
to 1.99% 



Vacuum Filtration 



6 lbs/£t /hr 
202 ft^ 

12 Ibs/ft^/hr 
49 ft^ 



Magnesium sludge 
CaCOj sludge 



Incineration 



T = 550"C 



22,298 lbs NVS 
remaining 



Carbonation 



9167 cfm 
161,435 lbs CO2 



Flue gas at 10% CO2 

++ 
100% Mg recovery 



133 



o 

o o 
.«2 S 



£o 
« o 

OO 

u 
a> 



f 






f 



/\ 



CM 

o_ 

o 



^ 



IS 

6 



/\ 



r~\ 



^ 



o 
a> c 

D O 

o 
o 



c 
o 



c 
CO 



£ 

2 >> 
CL a 

o ^ 
o = 
< PJ 



TTs 



■^ 



tn 
"o 

CO 



T^ 



£ 8 

3 a 
> Li- 



CM 



O 

en 
O 

a' 



^ 



CM 

o 
u 

ss 

o 

M— O 

t> CP 
to _3 . 



.c 
CM 



en 



CM 






/\ 



c 
o 



c: 
o 



5- 

VI 
J3 



in 
ro 

OJ" 
CM 



^ 






^ 



_\/ 



c 
_o 

o 
c 
o 

o 
o 

«J 
cc 



y\ 



cc: 



^ 



Pfs: 



a 

X^ If) 

5 9 



o 
o 
o 
in 






\/ 



a 



f 



e 

3 






■Jn 
0) 

g" >» 3 
O O TD 

■^ < o 

^- ?c- - 

yLcD 5 .2 > 

0) ro "^ .-= Si> 
tr CM OT3 >- 



u 
o 



o 

> 
o 

E 



o 
u 

E 

3 

"55 
V 

c 
o> 

o 
E 

E 



E 
o 
k. 
o» 
o 

o 



CM 

in 



ii. 



e 



o 
o 



CHAPTER 6 



COST 



6-1 Chemical Costs 

As o£ June 1976, magnesium sulfate was the most economical 
source of magnesium found by the suthor, and was available commer- 
cially at $120/ton. The cost of 300 mg/1 of Magnesium for treating 
a NSSC waste with an initial color of 5000 was $0,743/1000 gal. 
Lime was available commercially at $50/ton as of June 1976. The 
cost of 1250 mg/1 of lime to treat a NSSC waste with an initial 
color of 5000 was $0,260/1000 gal. 

The cost of the polymer 835A from American Cyanamid was $1.50/lb 
as of June 1976. The polymer dose was 3 mg/1 and added $0,037/1000 
gal to the cost of the lime -magnesium color removal process for NSSC 
waste. The CO2 gas was free since it would be taken from flue gas 
which is abundant at pulp and paper mills. The summarized chemical 
cost is presented in Table 6-1 and assumes no chemical recovery. 

6-2 Capital and Operation Costs 
All capital and operating costs were estimated from Liptak (1974) 



unless otherwise specified. The area required for settling treated 

2 

waste was 1020 ft per mgd of incoming waste. The capital cost of a 

settling basin was $18/ft^. The total capital cost for a 1, 5 and 10 



134 



135 



TABLE 6-1 

CHEMICAL COST TO TREAT A NSSC WASTE 
WITH AN INITIAL COLOR OF 5000 



Chemical $/1000 gal 



MgSO^ 0.743 

Ca(OH32 0.260 

835A 0.037 

Total 1 . 040 



136 



mgd waste flow would be $18,360, $91,800 and $183,600 respectively. 

The operating cost o£ vacuum filtration without heat treatment 
was estimated at $4/ton of dry solids for the sludge produced in the 
color removal process. This cost was $0,058/1000 gal on a unit flow 
basis and constant for any plant flow. The capital costs for Dorr- 
Oliver vacuum filters to concentrate the solids resulting from the 
treatment of a 1, 5 and 10 mgd flow were $100,000, $200,000 and 
$300,000 respectively. 

Approximately 3.4 tons of solids would be lost during incinera- 
tion for every million gallons of waste treated. The capital costs 
of equipment capable of incinerating the solids from treating a 1, 5 
and 10 mgd flow was estimated to be $455,000, $700,000 and $750,000 
respectively. The operating cost of the incineration process decrea- 
sed as the amount of material processed increased. Approximately 3.4, 
17.0 and 34.0 tons of solids would be incinerated when respective 
waste flows of 1, 5 and 10 mgd were treated. The cost per ton of dry 
solids incinerated was estimated to be $8.52, $3.80 and $3.50. The 
resulting unit costs per 1000 gallons of treated waste were calcula- 
ted to be $0,029, $0,013 and $0,012. 

The lime could be recovered from the carbonation process by 
passing the remaining slurry onto a vacuiun filter and then to a lime 
kiln. The CaCO_ would not dissolve in the carbonated slurry and 
would be available for recalcination. Only CaCO-r solids would remain 
after carbonation. After carbonation 14,094 lbs of CaCO^ sludge 
would be vacuum filtered per million gallons of treated waste. The 
filtration rate of the CaCO^ sludge was estimated to be twice that 



137 



o£ the sludge produced in the coagulation reaction. Vacuum filtra- 
tion o£ this sludge would cost $2.50/ton or $0,018/1000 gal on a 
unit flow basis. The costs for the capital equipment required to 
vacuum filter the CaCO„ sludge produced by treating a 1, 5 and 10 
mgd waste flow were estimated to be $40,000, $80,000 and $150,000 
respectively. 

The capital cost for the magnesium recovery phase was estima- 
ted by the author to be $2,500, $7,500 and $10,000 for waste flows 
of 1, 5 and 10 mgd respectively. No estimate, was made for the opera- 
ting cost of the magnesium recovery phase. These costs were covered 
in a miscellaneous estimate that -will be discussed later. 

The cost for recalcination at a 1 mgd plant was • estimated at 
$0.07/1000 gal for operating cost and $200,000 capital cost. A 5 mgd 
recalcination plant was estimated to cost $500,000 for capital . expen- 
ditures and $0.05/1000 gal for operating cost. The estimate for a 
10 mgd recalcination plant was $600,000 for capital expenditures and 
$0,035/1000 gal for operating cost. 

A final miscellaneous cost was estimated for the capital and 
operating costs for all other equipment necessary to install the 
color removal process. For a 1, 5 and 10 mgd effluent, the miscella- 
neous capital and operating costs respectively were $40,000 and 
$0,030/1000 gal, $100,000 and $0,020/1000 gal, and $150,000 and 
$0,015/1000 gal. 

6-3 System Costs 
In this section of Chapter 6 the appropriate cost of .different 



138 



color removal systems is presented. A complete system designates 
a color removal scheme using both Ca(OH) and magnesium recovery in 
the color removal process. All of the unit operations listed in 
Table 6-2 were used in the complete system as represented on the pro- 
cess flow chart in Figure 5.2. The other color removal systems 
differed only in the degree of chemical recovery. The costs of all 
systems considered are presented in Table 6-3. If Ca(QH32 recovery 
was eliminated, the costs associated with a vacuum filter for CaCO^ 
sludge and a lime kiln would be eliminated. If magnesium recovery 
was eliminated, the cost of vacuum filtration, incineration and car- 
bonation would be eliminated. However, the chemical costs associated 
with both of these systems would be increased. Complete chemical 
recovery was assumed in the cost calculations. The separate entries 
in Table 6-2 represent each unit operation used in the color removal 
process. 

The capital cost was calculated at 8% interest compounded annu- 
ally for 25 years. The unit cost was determined by dividing the 
amount of waste treated in a 25 year period into the capital cost. 
The unit cost per 1000 gallons of treated waste for each $100,000 of 
capital equipment required to treat a 1 mgd, 5 mgd and 10 mgd flow 
was $0.0761, $0.0152 and $0.0076. 

For a color removal process treating 1 mgd, the process cost 
siammary presented in Table 6-3 indicates that the most economical 
system required only magnesium recovery. The unit cost was $0,094 
per 1000 gallons. However, at 5 and 10 mgd, a process employing 
both magnesium and Ca(OH) recovery was the most economical system. 



139 



O 
O 
O 



T3 

e 



s 



03 
O 



bO 

O 
o 
o 



CM 
I 

M 
CQ 






O 



o 
o 
o 



o 

E- 



o 


CM 

o 


o 


r-H 

o 
o 


1— H 
O 


o 


r— 1 
(—1 

o 


1 — i 


o 


o 


o 


o 


o 


o 


o 


o 



o 

o 


o 
o 

• O 


o 
o 
o 


o 
o 
o 


o 
o 
o 


o 

o 
o 


o 

o 
o 


o 
o 


00 
f-H 


o 
o 
to 


o 

LO 


o 
I— 1 


o 

LO 

(— 1 


o 
o 


o 

LO 

1— 1 


I— H 

■feO- 



^T 


o 


r-- 


r-1 


IM 


\U 


LO 


lO 


t— 1 


lo 


o 


O 


r-H 


t^ 


1 — 1 


LO 


o 


o 


r-H 


o 


O 


o 


o 


OJ 


o 


o 


o 


o 


O 


o 


o 


o 



o 
o 

CO 


o 
o 
o 


o 

o 
o 


o 
o 

LO 


o 
o 
o 


c 
o 
o 


o 
o 
o 


o 
o 
^0 


1-H 


o 
o 


o 
o 


r-- 


o 

CO 


o 
o 

LO 


o 
o 

i-H 


1 — 1 



I— ( 
o 


o 


to 


o 
o 


o 
to 

o 


LO 

1— 1 


o 
to 
o 


o 

LO 


o 


o 


o 


o 


o 


o 


o 


o 



CO 

LO 

J o 


r-i 

o 


00 

1— 1 
o 


LO 

to 
o 


LO 

1— 1 
o 


to 

r-H 


o 


o 


o 


o 


o 


o 



oo to CO O o 

LO r-H r— I LO CN 

1 O O I O O O 



o o 



o o o 



LO 



o 


o 


o 


o 


o 


o 


o 


o 


^ 


o 


o 


o 


o 


o 


o 


\D 


to 


o 


o 


LO 


o 


o 


o 


CO 


CO 


o 


LO 


CN 


o 


o 


o 


LO 


I— t 


o 


LO 




'g- 


o 


■<* 


LO 




>— 1 


^ 






(Ni 




CO 









to 














to 












o 
















O 














y 
















U 














03 
















(3 














u 

1 
















u 










s 




■ (=: 










fi 






1 










o 




o 










o 






o 










•ri 




•H 










•H 






•H 










' -y 




+-> 








o 


J-1 






J-i 










Ri 




cd 




!/5 






oj 






Cd 




tfl 






u 


c 


^H 




3 






!h 


c 




!4 




3 






+-> 


o 


+-) 




o 






+-> 


o 




■P 




o 






1—1 


•H 


r-H 




(D 




■p 


i—H 


•H 




r— ( 




<D 






•H 


■M 


•H 


c 


c 


w 


in 


•H 


+J 




-H 


p! 


p; 


CO 




W) Ph 


ol 


Xo-, 


^H 


o3 


rH 


o 


(MP-. 


c« 


X Uh 


i-H 


cd 


r-H 




c 


^H 


fH 


•H 


1 — 1 


rt 


u 


P 


U 


u 




•H 


rH 


Cd 




•H E 


D 


e 


NS 


I— H 


■p 




•H 6 


o 


<D 


s 


t^ 


,-H 


-P 




■—I 3 


c 


> 3 




(U 


o 


p: 


■-H 3 


c 


> 


3 




CD 


O 




■P 3 


• H 


O 3 


0) 


o 


H 


o 


■P 3 


-H 


O 


p 


0) 


o 


E-. 


1—1 


■P o 


o 


o o 


e 


!/l 




•H 


■P O 


O 


u 


o 


B 


tfl 




03 


CD Oj 


c 


<D rt 


■H 


•H 




■p 


0} cd 


c 


0) 


03 


•H 


-H 




■l-J 

•H 


CO > 


1— 1 


Qi > 


hJ 


s 






CO > 


1 — 1 


oi 


> 


hJ 


s 




& 














0) 
















ni 














a, 
















u 














o 

















00 
LO 

1 o 


en 

CN 

O 


00 

t— 1 
1 o 


o 
o 


o 
to 

o 


LO 
CM 


o 


o 


o 


o 


o 


o 



140 



TABLE 6-3 

PROCESS COST SUMMARY IN $/1000 GALLONS 
OF NSSC WASTE WITH AN INITIAL COLOR OF 5000 



Costs 


Flow 
mgd 


Complete 
System 


No CaC0H)2 
Recovery 


No 
Magnesium 
Recovery 


No Ca(0H)2 

or 
Magnesium 
Recovery 


Capital 


1 


0.650 


0.468 


0.226 


0.044 




5 


0.255 


0.167 


0.117 


0.029 




10 


0.163 


0.106 


0.082 


0.025 


Operating 


1 


0.205 


0.117 


0.118 


0.030 




5 


0.159 


0.091 


0.088 


0.020 




10 


0.137 


0.084 


0.067 


0.015 


Chemical 


1 


0.037 


0.297 


0.780 


1.040 




5 


0.037 


0.297 


0.780 


1.040 




10 


0.037 


0.297 


0.780 


1.040 



Total 



1 


0.892 


0.882 


1.124 


1.114 


5 


0.451 


0.555 


0.985 


1.089 


10 


0.337 


0.487 


0.929 


1.080 



141 



The cost estimate for 5 and 10 mgd was $0,445/1000 gai and $0,322/1000 
gal respectively. 

In the color removal process, both a decreasing effluent color 
or an increasing plant flow will decrease unit cost. Most NSSC plant 
effluents range between 5 and' 10 mgd with colors varying from 2500 
to 5000. A cost estimate was made for the treatment of a NSSC waste 
with an initial color of 2500 for flows of 5 mgd and 10 mgd. These 
cost estimates were determined in the identical manner as those deter- 
mined for the stronger NSSC waste. They were based on the decreased 
solids loading produced from treating a weaker waste. These calcu- 
lations are summarized in Table 6-4, 

As can be '=:een from the summaries presented in Tables 6-3 and 
6-4, the unit cost in $/1000 gal will decrease with increasing plant 
flow and decreasing waste color. The unit cost estimates in $/1000 
gal for treating a 5 and 10 mgd flow of NSSC waste with an initial 
color of 2500 were calculated to be 0.371 and 0.273 $/1000 gal. 
When the color was assumed to be 5000, these estimates increased to 
0.451 and 0.337 $/1000 gal respectively. Comparison of these cost 
estimates indicates that flow has a greater effect on cost than does 
waste color. This implies that diluting a highly colored waste is 
not cost effective. Color streams should be concentrated whenever 
possible for the most economical treatment. A more economical unit 
cost was obtained from treating a 10 mgd waste at a color of 5000 
than a 5 mgd waste at a color of 2500. The most economical unit cost 
was, however, for the higher flow at the lower color. 

Approximately 20,000 gallons of NSSC waste are discharged for 



142 





Pi 




<t O 




is 




n CM 




C/3. 




H 


rt 


c/D oi 


1 


o o 


vO 


u -J 




o 


PJ 


:z u 


J 


o 


ea 


t-H tq 


< 


E- H 


H 


<: CO 




o; < 




m 3= 




D- 




o u 




CO 




H c/0 




KM 2 



.S 



03 
W) 

O 

o 
o 



o 
o 


o 


o 

LO 

o 


1— 1 
o 
o 


1 — 1 
1 — [ 

o 


o 


1 — 1 

rH 

o 


\0 

t— 1 


o 


o 


o 


o 


o 


o 


o 


o 



r- 


p- 


'* 


L/) 


LO 


00 


a-k 


to 


fO 


o 


l—f 


to 


1— i 


o 


t— 1 


r-- 


o 


o 


o 


o 


o 


1— 1 


o 


CNl 


o 


o 


o 


o 


o 


o 


o 


o 



o 






05 

b 
o 
o 



oj 
4-1 

o 



LO O LO o o o o 

LO o o o o o o 
vo o r-- o o o o 



p- LO to 
to CN LO 

rH CN vO 



o o o o 

r— I "^ O LO 
t— I ^O I— I 



O 
LO 

to 
1—1 



o 
o 


rsi 

o 


1— i 
CO 

o 


o 

o 


1— 1 
c 


o 


LO 

i-H 

o 


o 


o 


o 


o 


c 


o 


o 


o 


o 



r-- 


I-H 


'^t 


o 


o 


CN 


CTJ 


^ 


to 


.-H 


1 — I 


LO 


CM 


to 


r— 1 


t-- 


1 o 


o 


1 .o 


o 


o 


1 — 1 


o 


!0 


o 


c 


o 


o 


c 


o 


o 


o 



CO 


o 


r- 


o 


c 


o 


o 


lO 


fM 


o 


CTl 


o 


o 


o 


o 


CM 


00 


o 


to 


LO 


o 


c 


o 


r-- 


00 


o 


to 


r-. 


o 


o 


o 


CTi 


\D 


r^ 


to 




r-~ 


o 


o 


^ 




1 — 1 


LO 






LO 


1 — 1 













to 
















to 
















O 
















O 


















u 
















u 


















03 
















A 


















U 
















U 












C 






1 










R 






s 








■p 




O 






o 










O 






to 








I/) 




•H 






■ H 










•1-1 






•H 








o 




+J 






■P 










■P 






■P 








u 




oS 






o3 




m 






nj 






oi 




t/5 








U 


C 




U 




3 






^1 


C 




U 




3 




^ 




+-> 


o 




■(-> 




O 


■!-> 




P 


O 




■P 




o 




0) 




r-l 


•H 




i-H 




<D 


Ul 




1— 1 


• H 




,—1 




<u 


P 


p 


4-1 


•H 


+J 




•H 


c 


C 


o 




■H 


P 




•H 


C 


c 


iTl 


m 


tfl 


bO P-, 


03 


XPh 


1-H 


03 


u 


Wl ti. 


03 


Xtu 


t — 1 


03 


o 


X 


o 


c 


^H 


^H 




• H 


I-H 




!=; 




fH 


^ 




• H 


t— i 


u 


CO 


u 


■H S 


0) 


(D 


e 


X 


1 — 1 


(=; 


■H 


^ 


(U 





s 


t^ 


I-H 








■-H 3 


C 


> 


3 




<D 


o 


,-H 


3 


c 


> 


3 




<D 


1 — 1 


<D 


t-H 


+-> 3 


•H 


o 


3 


(D 


O 


•H 


■P 


3 


• H 


o 


3 


<D 


U 


o3 


P 


03 


+J o 


o 


o 


o 


e 


t/1 


■P 


P 


o 


O 


o 


o 


e 


1/5 


O 





+-) 


0) 03 


c 


a> 


03 


■ H 


■iH 


OJ 





OS 


c 


<D 


o3 


■H 


•H 


■ H 


I-H 


• iH 


W^ > 


1 — 1 


a 


> 


J 


^ 


f-l 


CO 


> 


I-H 


a 


> 


,-4 


S 


e 


& 


PL, 














(D 
















<D 


e 


o3 














ft 
















x: 


o 


u 














O 
















u 


CJ 



143 



every ton of NSSC product produced. The approximate cost to manufac- 
ture a ton of NSSC product is $200. The per cent cost increase of 
NSSC product if the magnesium color removal process was employed is 
given in Table 6-5. 



144 



TABLE 6-5 

NSSC PRODUCT COST INCREASE DUE TO COLOR REMOVAL 
BY MAGNESIUM COAGULATION 




CHAPTER 7 



CONCLUSIONS AND RECOMMENDATIONS 



7-1 Conclusions 
A process for removing in excess of 90% of the initial color 
in a NSSC waste has been developed based on laboratory experiments. 
The lime-magnesium color removal process mainly involves color 
removal by coagulation, .incineration of the color bodies and solu- 
bilizing the magnesium solids with protons from dissolved CO2 gas. 
The untreated NSSC waste was shown to have a significant acid 
strength. The amount of magnesium and Cam) 2 ^""^^^ ^° ^^^ ^^^^ 
waste in order to remove 90% of the color was directly proportional 
to the initial NSSC waste color. Increasing coagulant dose did not 
shift the optimum coagulant pH. When Ca(0H)2 was used for pH control 
instead of NaOH, the amount of magnesium required for 90% color 
removal from the NSSC waste was significantly decreased. Color 
removal achieved by magnesium precipitation in pH dependent and in 
pH independent chemical reactions demonstrated that magnesium was 
responsible for color removal in the lime -magnesium color removal 

process. 

The color bodies and the magnesium buffered the NSSC waste during 
the coagulation process. The magnesium ions first chelated the 
organic acids and increased the color, but eventually formed an 
insoluble organo-metallic precipitate that removed 90% of- the initial 



145 



146 



color. This was demonstrated by the increase in residual color for 
magnesium additions of 10-50 mg/1 to the waste solution. The color 
removal reaction proceeded by accepting the hydroxides from Ca(0H)2 
or NaOH. A precipitate was eventually produced that removed organic 
carbon, color and the magnesium from the NSSC waste. A 90% reduction 
in the initial color of the waste was accompanied by a 34% reduction 
in the organic carbon concentration. Approximately a 40% decrease 
of the acid strength of the untreated NSSC waste occurred during the 
color removal by magnesium coagulation. The jiajority of the reduc- 
tion in acid strength occurred in the weak acid range, pK = 9, which 
suggested that the weak acids were responsible for most of the color 
in the NSSC waste. The color was removed from the NSSC waste by a 
chemical reaction that involved color bodies, hydroxide and magnesium 
ions which resulted in the formation of an insoluble precipitate. The 
empirical formula of the precipitate was MgCOH). R, where R represen- 
ted the organic color bodies. 

Color removal from a NSSC waste was achieved by magnesium pre- 
cipitation in a pH dependent or a pH independent chemical reaction. 
Ca(OH)^ alone at the coagulation pH of the lime-magnesium color 
removal process did not remove any color from the NSSC waste. It 
was concluded that magnesium was responsible for color removal in 
the lime-magnesium color removal process. 

The sedimentation of the anionic colored floe was greatly assis- 
ted by the addition of an anionic polymer. The settling of the 
colored floe particles was controlled by polymer-floc bridging when 
the anionic polymer was present, not by electrostatic repulsion. 
Cationic polymers did not aid the settling of the colored floe. 



147 



The color removal process was reversible. Removal of the color 
bodies from the sludge hy incineration was required in order to reuse 
the magnesium for color removal. The optimum temperature of incinera- 
tion was 550°C. Increasing the incineration temperature from 550°C 
to 850°C did not remove any more color from the magnesium, but did 
remove some additional solids. 

The recovery experiments demonstrated that magnesium could be 
successfully reused in the lime -magnesium color removal process. In 
order to reuse the magnesium it was necessary^ to remove the color 
from the sludge by incineration and to dissolve the incinerated mag- 
nesium by carbonation. Approximately 93% of the original magnesium 
was recovered following three uses of the same magnesium in the lime- 
magnesiiim color removal process. There was no difference in the color 
reduction achieved by unused and recovered magnesium when the recovered 
magnesium was recycled with the solids remaining after carbonation. 

The unit cost of treating a NSSC waste by magnesium coagulation 
decreased as the volume of waste increased and the initial color of 
the waste decreased. However, because of the high capital expense of 
the equipment involved, the unit cost was more sensitive to the volume 
of waste treated than to the initial color of the waste. 

7-2 Recommendations 
Further research needs to be conducted on possible polymers that 
can serve as settling aids. Activated silica and starch are two 
anionic polymers that are more economical and may function as well as 
the partially hydrolyzed polyacrylamides. The NSSC waste should be 



148 



fractionated by gel filtration before and after magnesium coagulation 
to determine what fractions were removed during treatment. The frac- 
tions and functional' groups that contributed most to the color should 
be determined for the NSSC waste. Different pulp wastes could be 
treated by magnesium coagulation to determine if the color removal 
process can be adaptable to other pulp wastes. 

Possible areas of investigation for additional research would be 
the structure of the Mg(OH) (s) colloids and the rate of formation of 
Mg(0H3"^. Optimum rates of energy input into the rapid mix process 
should be determined using magnesium coagulation in different environ- 
ments. The fuel value of the sludge produced in the coagulation pro- 
cess should be determined and necessary experiments conducted to 
design an incinerator for the color removal process. The solids spe- 
cies controlling magnesium solubility in the carbonation process 
should be identified and the optimiun conditions for coagulant recycle 
determined. 

The lime-magnesium color removal process should be studied on 
a pilot plant scale. The design and operational parameters for this 
process can only be determined from such a study. The legal require- 
ments of the Water Quality Act and success of the lime-magnesiiom 
color removal process on a laboratory scale are good reasons to imple- 
ment this study. 



REFERENCES 



Alder, E.; Lundquist, K.; Miksche, E. "The Structures and Reactivity 
of Lignin." Advances in Chemistry Series 59 . R.F. Gould, ed. (1966). 

American Chemistry Society. "The Role of Adsorption of Hydrolyzed 
Aliiminum in the Kinetics of Coaglation-Adsorption from Aqueous 
Solutions." Advances in Chemistry Series 79- (1968). 

American Water Works Association Committee Report. "State of the 
Art of Coagulation." JAWWA , 63: 2: 99 (1971). 

Berger, H.F. U.S. Pat. 3,120,464 (Feb. 1964). 

Birkner, F.B.; Morgan, J.J. "Polymer Flocculation Kinetics of 
Dilute Colloidal Suspensions." JAWWA , 60: 175 (Feb. 1968). 

Black, A. P. "Full Scale Studies of the Magnesium Carbonate Water 
Treatment Process at Montgomery, Alabama and Melbourne, Florida." 
EPA Project #12120 HMZ (Sept. 1974). 

Black, A. P. "Some Applications of the Principles of Colloidal 
Behavior to Water Treatment." Proc. Rudolphs Res. Conf . , Rutgers 
State University, New Brunswick, N.J. (June 1960). 

Black, A. P.; Birkner, F.B.; Morgan, J.J. "Destabilization of Dilute 
Clay Suspensions with Labeled Polymers." JAWWA, Tl^. 175 (Dec. 1965). 

Black, A. P.; Christman, R.F. "Characteristics of Colored Surface 
Waters." JAWWA , 55_: 753 (1963a). 

Black, A. P.; Christman, R.F. "Chemical Characteristics of Fulvic 
Acids." JAWWA, 5£: 897 (1963b). 

Black, A. P.; Shuey, B.S.; Fleming, P.J. "Recovery of Calcium and 
Magnesium Values from Lime-Softening Sludges." JAWWA, 63_: 616 
(Oct. 1971). 

Black, A. P.; Singley, J.E.; Whittle, G.P.; Maulding, J.S. "Stoichio- 
metry of Coagulation of Color Causing Organic Compounds with 
Fe2(S0^)g." JAWWA , 5^: 1347 (Oct. 1963). 

Black, A. P.; Willems, D.G. "Electrophoretic Studies of Coagulation 
for the Removal of Organic Color." JAWWA , 53: 589 (1961). 

Canadian Pollution Abatement Research Program. "Color Removal from 
Biologically Treated Pulp and Paper Mill Effluents." CPAR Project 
Report 21 D-1 (1974). 



149 



150 



Christman, R.F.; Ghassemi, M. "Chemical Nature o£ Organic Color 
in Water." JAWWA , 58_: 723 (1966). 

Collins, J.W.; Webb, A. A.; Didwania, H.P.; Lueck, B.F. "Components 
of Wood Pulp Effluents." Env. Sci. ^ Tech. , Z: 371 (1969). 

Cornwell, D.A. "Recycling of Alum Used for Phosphate Removal in 
Domestic Wastewater Treatment." Ph.D. Dissertation, University of 
Florida (1975) . 

Davis, C. "The Effect of Sodium -Ion Concentration on the Removal of 
Color from Kraft Linerboard Mill Effluents." Paper presented at the 
20th anniversary meeting of Southeastern Tappi, Jacksonville, Florida 
(1972). 

Day, H.R.; Felbeck, G.T. "Production and Analyses of a Humic Acid- 
Like Exudate from the Aquatic Fungus Aureobasidium pullulans ." 
JAWWA , 66^: 8: 484 (Aug. 1964). 

Day, R.A., Jr.; Underwood; A.L. Quantitative Analysis . Prentice-Hall, 
Englewood Cliffs, N.J. (1967). 

Dollimore, D.; Horridge, T.A. "The Optimum Flocculant Concentration 
for Effective Flocculation of China Clay in Aqueous Suspensions." 
Water Research , Vol. b. Pergamon Press (1972). 

Domitar Limited Research Center. "Colour Removal from Biologically 
Treated Pulp and Paper Mill Effluents." CPAR Project Report 210-1 
(Mar. 1974). 

Dubose, A. "The Effect of Magnesium Concentration on Municipal 
Wastes." Ph.D. Dissertation, University of Florida (1973). 

Eidsness, F.A.; Black, A. P. "Carbonation of Water Softening Plant 
Sludge." JAWWA , 49^: 1343 (Oct. 1957). 

Fitzgerald, C.L.; Clemens, M.M.; Riley, P.B. "Coagulants for Waste 
Water Treatment." Chem. Eng. Progress , 66 : 36 (Jan. 1970). 

Freudenberg, K. "Analytical and Biochemical Background of a Consti- 
tutional Scheme of Lignin." Lignin Structure and Reactions, Advances 
in Chemistry Series 59 . R.F. Gould, ed. (1966). 

Gjessing, E.T.; Lee, G.F. "Fractionation of Organic Matter in 
Natural Waters on Sephadex Columns." Env. Sci. & Tech. , 7_: 631 (1967). 

Gjessing, E;T.; Samdal, J.E. "Humic Substances in Water and the 
Effect of Impoundment." JAWWA, 6£: 451 (1968). 

Gould, M. "Color Removal from Kraft Mill Effluent by an Improved 
Lime Process." Tappi , 56_: 79 (1973). 



151 



Guilledge, J.H.; O'Conner, J.T. "Removal of Arsenic (V) from Water 
by Adsorption on AlLiminum and Ferric Hydroxides." JAWWA, 65_: 8: 548 
(Aug. 1973). 

Hahn, H.N.; Stumra, w'. "Kinetics of Coagulation with Hydrolyzed 
Alum." Jour, of Coll. ^ Sci. , 288 : 134 (Sept. 1968). 

Hall, E.S.; Packham, R.F. "Coagulation of Organic Color with 
Hydrolyzing Coagulants." JAWWA, 57: 1149 (July 1965). 

Hannah, S.A.; Cohen, J.M.; Robeck, G.G. "Measurement of Floe 
Strength by Particle Counting." JAWWA, 59: 843 (July 1967). 

Harper, F.C. "Effect of Calcination Temperature on the Properties 
of Magnesium Oxides for Use in Magnesium Oxychloride Cements." 
Jour, of Applied Chem. , ]J: 1 (Jan. 1967). 

Herbet, A.J. "A Process for the Removal of Color from Bleached 
Kraft Effluents through Modification of the Chem-Recovery System." 
NCASI, Technical Bulletin No. 157 (June 1962). 

Interstate Paper Corp, "Color Removal from Kraft Pulping Effluent 
by Lime Addition." Water Pollution Control Research Series , 12040 
ENC (Dec. 1971). """ 

Jeffcoat, B.W.; Singley, J.E. "The Effect of Alum Concentration 
and Chemical Addition Times on Coagulation." JAWWA, 67_: 4: 177 
(Apr. 1975). 

Jensen, W.; Fremer, K.E.; Forss, K. "The Separation of Components 
Spent Sulfite Liquor." Tappi , 45: 122 (1964). 

Jobin, R.; Ghosh, M.M. "Effect of Buffer Intensity and Organic Matter 
on the Oxygenation of Ferrous Iron." JAWWA , 6£: 7: 590 (Sept. 1972). 

Kabeya, H.; Fujii, T.; Kubo, T.; Kimura, Y.; Urano, K. "Renovation 
of Pulp Mill Waste Water Adsorption Characteristics of Kraft Pulp 
Lignin on Activated Carbon." Kanipa Gikyoshi (Jap.), 26_: 3: 125. 
Chem. Abs. , 77_: 24541t (1972). 

Kawamura, S. "Coagulation Considerations." JAWWA , 65_: 6: 417 
(June 1973) . 

Kirk, T.E.; Brown, W. ; Cowling E.B. "Preparative Fractionation of 
Lignin by Gel -Permeation Chromatography." Biopolymers , 1_ (1969). 

Kirk, J.E.; Othmer, D.F. Encyclopedia of Chemical Technology , 2nd 
Edition, 12_. John Wiley § Sons, New York (1967). 



152 



Kotoh, S.; Kimura, Y. "Study of Renovation of Pulp Mill Waste Water 
by Treatment with Fly Ash." Jour. Jap. Tappi , 25^: 4: 168 (1972). 
Abs. Bull. Inst. Paper Chem. , 42: 8176 (1972). 

LaMer, V.K. "Coagulation vs. Flocculation of Colloidal Dispersions 
by High Polymers." Principles fi Applications of Water Chemistry. 
(4th Rudolphs Conf.). Faust, S.D.; Hunter, J.V. Principles and 
Applications of Water Chemistry , Wiley Interscience (1967) . 

Langelier, W.F.; Ludwig, N.F. "Mechanism of Flocculation in Clarifi- 
cation of Turbid Waters." JAWWA, 41: 163 (Feb. 1949). 

Leszczynski, C. "Decolorization of Kraft Mill Effluents." 

Prezegl. Papier (Pol.), 28_: 3: 88 (1972). Abs. Bull. Inst. Paper Chem. 

43_: 4154 (1972). 

Luner, P.; Dence, C. "Mechanisms of Color Removal in the Treatment 
of Pulping and Bleaching Effluents with Lime." Vol. I. Treatment of 
Caustic Extraction Stage Bleaching - Effluent . Tech Bulletin 239 
(July 1970). 

Luner, P.; Dence, C; Bennett, 0.; Ota, M. "The Mechanism of Color 
Removal in the Treatment of Pulp and Bleaching Effluents with Lime." 
NCASI, Technical Bulletin No. 242 (Dec. 1970). 

Mangravite," F.J., Jr.; Buzzell, T.D.; Cassell, E.A.; Matijevic, E.; 
Saxton, G.B. "Removal of Humic Acid by Coagulation and Microflota- 
tion." JAWWA, 6^: 288 (Feb. 1975). 

Matijevic, E.; Janauer, G.E.; Kerker, M. "Reversal of Charge of 
Lyphobic Colloids by Hydrolyzed Metal Ions." Jour, of Coll. ^ Sci. , 
19_: 333 (1964) . 

Maulding, J.S.; Harris, R.H. "Effect of Ionic Environment and 
Temperature on Coagulation of Color-Causing Organic Compounds with 
Ferric Sulfate." JAWWA , 60: 460 (1968). 

Michaels, A.S. "Aggregation of Suspensions by Polyelectrolytes." 
Ind. S Eng. Chem. , 46 : 1485 (July 1954) . 

Midwood, R.B.; Felbeck, G.T. "Analysis of Yellow Organic from Fresh 
Water." JAWWA , 60: 352 (1968). 

Mohtaoi, M.F.; Roa, P.N. "Effect of Temperature on Flocculation of 
Aqueous Dispersions." Water Research , 17 . Pergamon Press (1973). 

Morrow, J.J..; Rausch, E.G. "Colloidal Destabilization with Cationic 
Polyelectrolytes as Affected by Velocity Gradients." JAWWA , 66 : 11: 
646 (Nov. 1974). 



153 



Narkis H • Rebhum. "The Mechanism of Flocculation Processes in the 
Presence of Humic Substances." JAWWA, 67_: 2: 101 (Feb. 1975]. 

NCASI Tech. Bulletins 249 and 212. "Surface Properties of Hydrogels 
Resulting from Treatment of Pulp and Papermill Effluents." Parts 
I and II (Sept. 1971). 

NCASI Tech Bulletin 273. "Studies on Adsorption of Spent Chlorina- 
tion and Spent Caustic Extraction Stage Liquor Color and Organic 
Carbon on Activated Carbon." (1974). 

Oldham, W.F.; Gloyna, E.F. "Effect of Colored Organics on Iron 
Removal." JAWWA , 61_: 611 (1969). 

Olthof, M.G.; Eckenfelder, W.W., Jr. "A Laboratory Study of Color 
Removal from Pulp and Paper Wastewaters by Coagulation." Tappi, 
57^: 8: 55 (1974). 

O'Melia C R "Coagulation and Flocculation." Physiochemic al Processes 
for Water Quality Control . W.J. Weber, Jr., ed. Wiley Interscience, 
New York (1972) . 

Osipow, L.I. Surface Chemistry Theory and Indust rial Application. 
Robert E. Krieger Publishing Co., Huntington, New York (1972). 

Packham, R.F. "Some Studies of Coagulation of Dispersed Clays with 
Hydrolyzing Salts." Jour, of Coll. § Sci. , 20_: 81 (1965). 

Packham, R.F. "Studies of Organic Color in Natural Water." 
Proc. Soc. Water Treat. Exam. , 13^: 316 (1964) . 

Posselt, H.S.; Anderson, F.J.; Weber, W.J., Jr. "Cation Sorption on 
Colloidal Hydrous Manganese Dioxide." Env. Sci. 5 Tech. , 2_: 1087 
(1968b) . 

Posselt, H.S.; Reides, A.H.; Weber, W.J., Jr. "Coagulation of 
Colloidal Hydrous Manganese Dioxide." JAWWA , 60_: 48 (1968a). 

Predali, J.J.; Cases, J.M. "Ze_at, Potential of Magnesium Carbonates 
in Inorganic Electrolytes." Jour, of Coll. S Inter. Sci. , 45: 3 
(Dec. 1973). 

Pruill, E.R.S. "Color Removal and Sludge Disposal for Kraft Mill 
Effluents." Paper Trade Journal , (Aug. 1974). 

Ragunathan, P.; Cleasby, J.L.; Cerwick, J. A. "Coagulation, Cake 
Filtration and Filterability." JAWWA, 65: 3: 202 (Mar. 1973). 

Rapson, B.; Sullivan, D.P.; Brothers, J.S. "The NRSF Sea Water-Lime 
Clarification Process for Kraft Mill Effluents." Presented at the 60th 
Annual Meeting of Technical Section C.P.P.A., Montreal, Canada (1971). 



154 



Robinson, C.N.; Correl, N., Jr. "Polyelectrolytes as Primary 
Coagulants for Potable Water Systems." JAWWA , 66^: 4: 252 (April 1974]. 

Saville, T. "The Nature of Organic Color in Water." Jour, of New Eng. 
Water Works , 31_: 79 (1917) . 

Schenk, J.E.; Weber, W.J., Jr. "Chemical Interactions of Dissolved 
Silica with Iron II and III." JAWWA, 60_: 199 (1968). 

Schnitzer, M. "Metal Organic Matter Interactions in Soils and 
Waters." Organic Compounds in Aquatic Environments . S.D. Faust, J.V. 
Hunter, eds. Marcel Dekker Inc., New York (1971). 

Schott, H. "Deflocculation of Swelling Clays by Nonionic and Anionic 
Detergents." Jour. Coll. ^ Inter. Sci. , 26^: 133 (1968). 

Sethuraman, V.V.; Raymahashoy, C.B. "Color Removal by Clays." 
JAWWA , 13: 9: 1139 (Dec. 1975). 

Shapiro, J. "Effect of Yellow Organic Acids on Iron and Other 
Metals in Water." JAWWA , 56: 1062 (1958). 

Singley, J.E.; Black, A. P. "Hydrolysis Products of Iron (III)." 
JAWWA , 59: 1549 (Dec. 1967). 

Smith, S.E.; Christman, R.F. "Coagulation of Pulping Wastes for the 
Removal of Color." JWPCF, 41: 2: 222 (Feb. 1969). 

Spruill, E.L. "Color Removal and Sludge Recovery from Total Mill 
Effluent." Tappi , 56: 98 (1973). 

Stumm, W. "Metal Ions in Aqueous Solutions." Principles and 
Applications of Water Chemistry , S.D. Faust, J.V. Hunter, eds. 
Wiley, New York. 520 (1967). 

Stumm, W.; Huper, H.; Charaplin, R.L. "Formation of Polysilicates 
as Determined by Coagulation Effects." Env. Sci. ^ Tech. , l_: 221 
(1967). 

Stumm, W. ; Lee, G.F. "Oxygenation of Ferrous Iron." Ind . Eng . Chem . , 
53: 143 (1961). 

Stumm, W.; Morgan, J. "Chemical Aspects of Coagulation." JAWWA , 
54: 971 (1962). . 

Stumm, W.; O'Melia, C.R. "Stoichiometry of Coagulation." JAWWA , 
60: 514 (1968). 

Swanson, J.W.; Dugal, H.B.; Buchanan, M.A.; Dickey, E.E. "Kraft 
Effluent Color Characterization Before and After S'toichiometric 
Lime Treatment." Env. Prot. Agency Series, EPA R2-73-141 (1973). 



155 



Taflin, CO.; Weber N.F.; Kramer, A.S.; Whitaker, J. "Minneapolis 
Keeps on Trucking." Water ^ Wastes Eng. , Reprint (May 1975). 

Tajera, N.E.; Davies, M.W.,Jr. "Removal of Color and Organic Matter 
from Kraft Mill Caustic Extraction Waste by Coagulation." Tappi , 
53_: 10 (Oct. 1970). 

Taylor, J.S.; Zoltek, J, Jr. Unpublished Research Data, Univ. of Fla., 
Gainesville, Florida (1974). 

Tenney, M.W.; Stumm, W. "Chemical Flocculation of Microorganisms in 
Biological Waste Treatment." JWPCF , 87: 1370 (1975). 

Thompson, C.G.; Singley, J.E.; Black, A. P. "Magnesium Carbonate - 
A Recycled Coagulant." JAMA , 6£: 1: 11 (1972a). 

Thompson, C.G.; Singley, J.E.; Black, A. P. "Magnesium Carbonate - 
A Recycled Coagulant - II." JAWWA, 64: 2: 95 (1972b). 

Zettlemayer, A.C.; Micale, F.T.; Dole, L.R. "Surface Properties of 
Hydrogels Resulting from Treatment of Pulp and Paper." NCASI Tech. 
Bulletin 225 (Dec. 1968). 

Zoganathan, P.; Maier, W.J. "Some Surface Chemical Aspects in 
Turbidity Removal by Sand Filtration." JAWWA, 67: 6: 336 (1975). 



Biographical Sketch 

James Sherman Taylor was born August 24, 1941, in Miami, 
Oklahoma. He graduated from secondary school at Miami High School 
in Miami, Oklahoma and attended Oklahoma State University in 
Stillwater, Oklahoma on a football scholarship. He received the 
degree of Bachelor of Science in Industrial Engineering and Manage- 
ment in August 1965, from Oklahoma State University. 

He accepted a position as process and industrial engineer with 
3M Company in Hastings, Minnesota from September 1966 until May 
1967. He then worked as a research engineer on the Saturn program 
in Cape Canaveral, Florida for one year. In June 1968, he accepted 
a position as a senior engineer with Radiation Incorporated in 
Melbourne, Florida. He left Radiation in January 1971 to pursue a 
Masters degree in engineering at the University of Florida in Gaines- 
ville, Florida. In June 1972 he received a Master of Engineering 
degree in Environmental Engineering from the University of Florida. 
He entered the Ph.D. program in the Department of Environmental 
Engineering Sciences at the University of Florida in June 1972. He 
became a registered engineer in the state of Florida in April 1974. 

He is married to Janet Louise Taylor, formerly of Melbourne, 
Florida, and has two children, James Sherman Taylor II, age 11, and 
Briton Ashley Taylor, age 5. The author presently is an Assistant 
Professor of Environmental Engineering Sciences at the Florida 
Institute of Technology in Melbourne, Florida. 



156 



I certify that I have read this study and that in my opinion it 
conforms to acceptable standards of scholarly presentation and is 
fully adequate, in scope and quality, as a dissertation for the degree 
of Doctor of Philosophy. 




k 



Zoltek, 'jy. , Chairmiin 
Associate Professor of 
Environmental Engineering Sciences 



I certify that I have read this study and that in my opinion it 
conforms to acceptable standards of scholarly presentation and is 
fully adequate, in scope and quality, as a dissertation Lor tne degree 
of Doctor of Philosophy. 



/ 



■U.^tv---^^^ 



r. 



( 



•cc -» 



T. deS. Furman 

Professor of 

Environmental Engineering Sciences 



I certify that I have read this study and that in my opinion it 
conforms to acceptable standards of scholarly presentation and is 
fully adequate, in scope and quality, as a dissertation for the degree 
of Doctor of Philosophy. 





:>ir'. Edward Sing ley . 
Professor of 
Environmental Engineering Sciences 



I certify that I have read this study and that in my opinion it 
conforms to acceptable standards of scholarly presentation and is 
fully adequate, in scope and quality, as a dissertation for the degree 
of Doctor of Philosophy. 




Ellis D. Verink, Jr. / 
Professor and Chairman of 
Material Science and Engijieering 



This dissertation was submitted to the Graduate Faculty o£ the 
College of Engineering and to the Graduate Council, and was 
accepted as partial fulfillment of the requirements for the 
degree of Doctor of Philosophy. 

August 1976 



■( ;'<^-e.V''''<- 



// CU- 



Dean, College of Engineering 



Dean, Graduate School